Electrostatics 2003 (Institute of Physics Conference Series)

© 2004 by Taylor & Francis Group, LLC

288

Figure 3. Current versus electric field.

Figure 4. Discharge current versus time (U0=0 m/s and E=7.6 kV/cm).

3.2 Influence of the free air stream velocity U0 The local current density in ith direction Ji depends on the local magnitudes of the free charge density ρc, the ion diffusion coefficient D, the ion mobility µ, the electric field Ei and the convective velocity ui in ith direction. It may be expressed as: (1) In classical electrohydrodynamical theory, the diffusive and convective terms are usually neglected. This simplification is justified because these terms are largely lesser than the Coulomb drift term. However in aeronautic applications, convective velocity is high and it can not be neglected. For instance, in our experiments and considering the ion mobility µ≈2.0 10–4 m2/Vs, the Coulomb drift velocity µEi is about 150 m/s while ui is comparable to U0≈30 m/s. Figure 5 presents typical curves of current versus electric field for U0 equal to 0 and 30 m/s. It shows that the discharge current and the corona starting voltage increases with U0. In Figure 6, square symbols, the electric field E is fixed at 7.5 kV/cm, U0 at 0 m/s and the current is about 0.4 mA/m. If U0 is increased we can observe that the discharge current increases almost linearly with the free air stream velocity as predicted by Eq. (1). Moreover, below 12 m/s the discharge is “high spot” while above 12 m/s it becomes “glow”. This means that the airflow favours the “glow discharge” in spite of the “high spot” type discharge. The same behavior is obtained with E equal to 8 and 8.25 kV/cm. In order to increase the convective term ui, we have inclined down the leading edge of the plate to an incidence angle α of 10°. As in Figure 5, Figure 7 presents the current-electric field characteristics for different U0 values. It shows that the slope of the E-I characteristic increases with U0. Taking into account Eq. 1 and if we consider that the cross-sectional area of the discharge is constant, then the slope of these characteristics is ρc×µ. Then the slope increase may be associated to an increase of the space charge density ρc with U0. Figure 8 presents the current as a function of the free airstream velocity for a equal to 0 and 10 degrees and E=8 kV/cm. As previously indicated by Figure 6 for α=0, it shows that the discharge current increases with U0. The effect is higher when the plate is inclined because the convective term ui is higher. On the other hand, if the air stream flows in the opposite direction to the current, the convective term ui of Eq. 1 should diminish the current value because it is negative. However when the flat plate is reversed in the wind tunnel loop, the current does not decrease but increases slightly.

© 2004 by Taylor & Francis Group, LLC

289 This result shows that the influence of the air flow is not restricted to the convective effect. It indicates that the phenomena of charge generation, charge distribution, charge drift and convection are strongly coupled. Figure 9 presents the alternative component of i (t) at 30 m/s for α=10°. Compared to Figure 4, Figure 9 shows that the airflow largely increases the peak repetition and modifies the discharge spectra. This confirms that the airflow modifies considerably the discharge.

Figure 5. Current versus electric field for 2 air stream velocity values,

Figure 6. Current versus air stream velocity for 3 electric field values.

Figure 7. Current versus electric field for 4 airstream velocities with α=10°.

Figure 8. Current versus free air stream velocity for E=8 kV/cm.

Figure 9. Alternating component of the current (ImeanH”300 µA) versus time for U0=30 m/s and α=10°.

Figure 10. Current versus electric field for different air pressure values (in 105 Pa).

© 2004 by Taylor & Francis Group, LLC

290 3.3. Influence of the air pressure P One parameter which may change in aeronautic applications for example is gas pressure. This is the reason why we studied the discharge behaviour in a large range of pressures (102
Léger L, Moreau E, Artana G, Touchard G 2001 J.of Electrostatics 50–51 448–454 Léger L, Moreau E and Touchard G 2002 IEEE Trans.on Ind. Appl. 38 1478–1485 Léger L, Moreau E and Touchard G 2002 1st AIAA Flow Control Conf. paper #2833 Artana G, D’Adamo J, Moreau E, Touchard G 2002 AIAA Journal 40 1773–1779 Artana G, Diprimio G, Desimone G, Moreau E, Touchard G 2001 4th AIAA Weakly ionized Gases Conf., paper #3056 Moreau E, Léger L and Touchard G 2002 IEEE—CEIDP 272–278 Artana G, Desimone G and Touchard G 1999 Electrostatics Cambridge 147–152 Louste C, Moreau E and Touchard G 2003 Electrostatics Edimbitrg Louste C, Moreau E and Touchard G 2002 IEEE—CEIDP 822–826

© 2004 by Taylor & Francis Group, LLC

291

Optimized geometry of a corona electrode arrangement for water ozonization Ilie Suarasan,1 Roman Morar,1 Lucia Ghizdavu,2 Luciae Dascalescu3 High Intensity Electric Fields Research Laboratory, Technical University, 15 C-Daicoviciu Street, 3400 Cluj-Napoca, Romania 2 Chemical Department, Babes-Bolyai University, 3400 Cluj-Napoca, Romania 3 Electronics and Electrostatics Research Unit, LAII-ESIP, UPRES-EA 1219 University Institute of Technology, 4 Varsovie Ave., 16021 Angouleme, France 1

Abstract. Barrier electric discharges are commonly used for generating the ozone employed for water treatment. Previous studies demonstrated the feasibility of using corona discharge for potable or waste water ozonization. The present paper aims at analysing the effect of corona electrode geometry on the efficiency of ozone generation (amount of ozone/consumed energy). The experiments were carried out with three types of high-voltage supplies energizing a test cell consisting of a needle-type corona electrode, facing an horizontal plate (collecting) electrode, and a Petri dish, containing the treated liquid: potable, mono- or bi-distiled water. The ozone measurements were correlated to the current-voltage characteristics of the electrode system, for various values of the following parameters: interelectrode spacings, distance between adjacent needles of the corona electrode p, high voltage level U. A special device carrying a current probe was employed for measuring the distribution of corona current density at the surface of the collecting electrode. The measurements showed that the most effective ozone generation is obtained for a s/p ratio slightly greater that 1.

1. Introduction During the last 20 years, various novel methods of ozone generation [1–5] have been studied as competitors for the classical Siemens water treatment reactors, based on dielectric barrier discharges [6–9]. Several promising results have been published by those who investigated the production of active species such as OH, H2O2, and O3 in water and aqueous solution using the so-called pulsed-streamer corona discharge [10–13]. A combination of air stripping and pulsed corona has also been tested [14–15]. Other research groups have studied ozone generation in devices where DC corona is created between an electrode and the surface of a liquid [16]. In a previous work [17], the authors attempted to evaluate AC corona discharges from multipoint high-voltage electrodes above a water film as means for direct ozonization of the respective liquid. The reported results showed that the efficiency of ozone generation depends 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; (v) presence of a dielectric barrier between the collecting electrode and the liquid. The aim of the present paper is to analyse more in depth the effect of corona electrode

© 2004 by Taylor & Francis Group, LLC

292 geometry on ozone generation, and this for various types of high-voltage supplies: AC, DC (full-or half-wave rectified). The study was suggested by previously reported results that show that the efficiency of ozonization is strongly related to the characteristics of the corona discharge, mainly to the intensity of the corona current. Therefore, an attempt was made to find the electrode arrangement capable to ensure a high and uniform density of the corona current in the air-gap above the liquid to be treated. 2. Experimental procedure The study was carried out on an experimental device composed of an “active” high-voltage electrode and a grounded plate electrode (Fig. 1). The liquid to be treated (volume: 10 to 70 ml) was contained in a Petri glass vessel of inner diameter Φ=95 mm. Two types of “active” electrodes were tested. The first consisted of one to three metallic points, disposed at various separations (p=1–2.5–5–7.5–10 mm) on a cylindrical metallic support. The second, not presented in the figure, consisted of a multi-point (brush-type) arrangement, the distance between adjacent spikes being 1 mm. The tungsten wire segments of 150 ˜m 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 40 mm. The 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 electrode system was fed from a continuously adjustable custom-designed 0–25 kV, 5 mA, 50 Hz, high voltage supply, associated—in some experiments—with either a half- or a Miwave solid state rectifier. The device that was employed for the measurement of the corona current distribution generated by the various electrode arrangements is presented in Fig. 2. A current probe, consisting of the exposed end of an enameled copper wire (diameter: 0.4 mm; the thickness of the enamel coating was estimated to be >10 µm), was embedded in the center of the collecting electrode (a polished aluminum plate, 140 mm×200 mm, with rounded edges, so that to avoid “parasitic” discharges). The probe was connected to a nano-ammeter. The distance between the tips of the needles and the collecting electrode facing them

Fig. 1. Schematic representation of the experimental device; 1—AC high voltage power supply, 25 kV, 5 mA, 50 Hz+half- or full-wave rectifier; 2—Petri vessel containing the liquid to be treated; 3—active electrode; 4—treated liquid; 5—metallic contact between the treated liquid and the grounded electrode; 6—grounded plate electrode (all the dimensions are in mm).

© 2004 by Taylor & Francis Group, LLC

293

Fig. 2. Experimental device for the measurement of the corona current density at the surface of the collecting electrode.

could be continuously adjusted from 5 to 50 mm. The multiple-needle electrode was attached to an X-Y positioning system. The Y-axis is defined by the projections of a row of needles on the surface of the collecting electrode. For a given inter-electrode distance and a fixed position of the collecting plate, the needles were translated along two orthogonal directions, so that the current probe could scan a 75 mm×75 mm rectangle. 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 (70 ml), which was then acidified by adding 5 ml of H2SO4 solution (2N), for the delivery of iodine from the complex of KI3. The delivered iodide atoms were visualized (blue coloration) by using 1 ml of amidon solution (0.5%), and their quantity was determined by titration with a Na2S2O3 solution (0.00 IN). The details of the method have been given in [17]. 3. Results In a first set of experiments, carried out on the device presented in Fig. 1, current-voltage characteristics were obtained for various electrode configurations. An example is shown in Fig. 3. The curves in Fig. 4, obtained during this same set of experiments illustrate the variation

Fig. 3. Current-voltage characteristics of the air gap between a two-needle corona electrode and the liquid, at two values of the distance p between the corona emitting points. The “treated” liquid is either connected to the earth (case I) or at floating potential (case II) with respect to the grounded electrode

© 2004 by Taylor & Francis Group, LLC

294

Fig. 4. Variation of the total corona current with the distance between adjacent points.

Fig. 5. Ozone quantity as a function of the inter-points distance, for a discharge gap s=20 mm, an applied high-voltage U=20 kV, 50 Hz, and a discharge exposure duration of 30 seconds.

of the corona current with the distance p between two adjacent points of discharge. The relation between this distance p and the quantity of ozone detected in the mass of the liquid is shown in Fig. 5. The efficiency of ozone generation, defined as the ratio between the quantity of O3 in water and the intensity of the corona current is calculated in Fig. 6, for various p. The results given in Figs. 3 to 6 were obtained with the “treated” liquid either connected to the earth (case I) or at floating potential (case II) with respect to the grounded electrode The second set of experiments, carried out on the experimental devices presented in Fig. 2 enabled the measurement of the corona current density at the surface of the collecting electrode, for various arrangements. Two sets of typical curves are displayed in Figs. 7 and 8.

© 2004 by Taylor & Francis Group, LLC

295

Fig. 6. Ozone generation efficiency as function of the distance p between adjacent points of the corona electrode, for a discharge gap s=20 mm, an applied high-voltage 20 kV, 50 Hz, and an exposure time of 60 s.

Fig. 7. Distribution of the measured current density at the surface of the collecting electrode for three values of the air-gap s (x is measured from the central axis of the electrode system in a direction that is perpendicular to the plane containing the three coronating points, p=20 mm).

Fig. 8. Distribution of the measured current density at the surface of the collecting electrode for three pairs of values (s, U); for p=20 mm. Note that y is measured from the central axis of the electrode system in the direction defined by the plane containing the three corona-emitting points. Thus, the central of the three points corresponds to y=0 mm, and an adjacent point is at y=20 mm. The current density is maximum in the axis of the points and null approximately midway between them.

© 2004 by Taylor & Francis Group, LLC

296 4. Discussion The correlation between the intensity of the electric current and the amount of ozone is welldocumented [7]. This justifies the interest of current-voltage characteristics in Fig. 3. These curves indicate that, at given voltage, the current is slightly higher and the corresponding quantity of ozone is likely to be larger when the corona emitting points are less close to each other. The characteristics displayed in Fig. 5 clearly confirm this prediction, for both cases under study (with the liquid grounded or at floating potential) and for both 2- and 3- point electrode models. The optimum in terms of efficiency (quantity of ozone obtained for 1 mA of corona current) is obtained for p smaller than the discharge gap s, as shown in Fig. 6. This finding is not completely explained by the aspect of the I(p) curves in Fig. 4, and prompted the second set of experiments, regarding the spatial distribution of the corona discharge. The aspect of the curves given in Figs 7 and 8 indicate that the distribution of the corona current density is not uniform at the surface of the liquid. By reducing the distance between the points, the volume occupied by the corona discharge diminishes and may explain the diminution of the quantity of ozone, in spite of the fact that, as shown in Fig. 4, the total current is quasi-constant. The curves of current density distribution in Figs. 7 and 8 diverge significantly from those predicted by the widely-accepted Wartburg law. The explanation resides in the presence of a metallic support in the proximity of and connected at the same high-voltage supply as the coronaemitting points. The effect of this support is to intensify the field in a larger region of the discharge gap and hence to expand the region where ozonization takes place. The development of electrode systems for the direct ozonization of liquids should take into account these observations and the results of further research on other electrode designs. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

Yoshida K and Tagashira H 1986 Mem. Kitami Inst.Technol. 1611 Okazaki S, Sugimutsu H, Niwa H, Kogoma M, Moriwaki T and Inomata T 1988 Ozone Sci. Eng. 10 137 Braun D, Küchler U and Pietsch G 1991 J. Phys. D: Appl. Phys 24 564 El-Mohandes M T, Ushiroda S, Kajita S, Kondo Y and Horii K 1992 IEEE Trans. Ind. Appl. 1 1567 Hegeler F and Akiyama H 1997 IEEE Trans. Plasma Sci. 25 Eliasson B, Hirth M and Kogelschatz U 1987 J. Phys. D: Appl. Phys. 20 1421 Kogelschatz U and Eliasson B 1995 Ozone generation and applications. In: Chang, J.S., Kelly, A., Crowley, J.M. (eds.), Handbook of Electrostatic Processes (New York: Marcel Dekker) 585 Kitayama J and Kuzumoto M 1997 J. Phys. D: Appl. Phys. 30 2453 Yagi S and Tanaka M 1979 J Phys D: Appl. Phys. 12 1509 Wilberg D M, Lang P S, Hochemer R H, Kratel A and Hoffmann M R 1996 Environ. Sci. Technol. 30 2526 Clements J S, Sato M and Davis R H 1987 IEEE Trans Ind. Appl. 23 224 Sato M, Ohgiyama T and Clements J S 1996 IEEE Trans Ind. Appl., 32 106 Sun B, Sato M, Harano A and Clements J S 1998 J. Electrostatics, 43 115 Ai-Arainy A A, Jayaram S. and Cross, J.D. 1996 Conf. Rec. ICDL (Rome, Italy) 427 Lubucki P, Jayaram S, Cross J D and Al-Arainy A A 1996 Ann. Rep. CEIDP (San Francisco) 730 Goheen S C, Mong G M, Pillay G and Camaioni D M 1994 First Int. Conf. Advanced Oxidation Technology for Water and Air Remediation, London, Ontario, 1994, pp. 83–84. Suarasan I, Ghizdavu L, Ghizdavu I, Budu S and Dascalescu L 2002 J. Electrostat. 54 207

© 2004 by Taylor & Francis Group, LLC

297

Characterisation of ESD waveform and peak current from charged printed wiring board to ESD hand tools T Kalliohaka and J Paasi VTT Industrial Systems, PO Box 1306, FIN-33101 Tampere, Finland Abstract. Electrostatic discharge (ESD) protective hand tools are used in the manufacturing of electronics in order to protect sensitive electronics from ESD failures. A hand tool is commonly made ESD protective by adding resistance into the handle of the tool. This added resistance reduces the peak current of human body type ESD, greatly reducing the risk of ESD failure. The situation becomes different when the discharge is from a charged device to the hand tool. In this situation the resistance between the ESD hand tool and the charged device can be almost zero leading to the most severe form of ESD when two metal objects at different electrostatic potentials come into contact. In this paper we investigate both charged device model (CDM) and human body model (HBM) type of ESD. A charged printed wiring board (PWB) is discharged by the tip of ESD hand tool in the CDM test, and a charged person is discharged to a PWB in the HBM test. Different kinds of hand tools are compared: metallic pliers with electrostatic dissipative handle, metallic tweezers with carbon fibre tip and carbon fibre tweezers with non-metallic parts etc.

1. Introduction Choosing a correct hand tool is important when handling very BSD-sensitive devices (ESDS). In this paper we investigate different kinds of hand tools in both CDM and HBM types of discharges. A charged PWB with ESDS components must be handled very carefully with hand tools. In a CDM-type of discharge charge transfer from the PWB through a pin of ESDS is very rapid and may damage the ESDS. Resistance between an ESDS pin and a hand tool and the capacitance of the PWB as well as the hand tool determine the peak current of the CDM ESD pulse. In addition to the resistance and capacitance of the ESDS—hand tool system, the total amount of transferred charge is influenced also by the capacitance of the human body and the resistance from the tip of a hand tool through the body to ground. 2. Test set-up 2.1 CDM ESD test The tests were done using an unpopulated 4-layer PWB with FR4 insulator. The layers of the PWB were connected so that the PWB has maximum capacitance. Dimensions of the PWB were 130×200×1.3 mm. The bottom surface of the PWB was connected to ground and a 16 mm long, 0.6 mm diameter, sharp point needle was attached on top of the PWB with

© 2004 by Taylor & Francis Group, LLC

298 silver loaded solder. A high frequency current transformer was attached so that the needle goes through the current transformer. The current transformer was Tektronix CT1. In the measurement, the PWB was at first, charged to 100 volts. Then a person touched the tip of the needle with the hand tool tip to achieve a discharge from the PWB to the hand tool. Current waveform was measured with a high bandwidth oscilloscope, Agilent Infmiium 54845A having 1.5 GHz bandwidth and 8 Gs/s maximum sample rate. The PWB layer structure and the sketch of the measurement system is presented in Fig. 1. The capacitance of the PWB was measured with Fluke 79III multimeter as 9.4 nF. The charge of the PWB was measured with Monroe Nanocoulombmeter 284. The measured charge of the PWB was 930 nC when the PWB was powered up to 100 V. The measured capacitance value was very close to that (9.3 nF) calculated from the measured charge. In tests, the person was holding a hand tool in the left hand and grounded by a wrist wrap in the right hand. Resistance-to-ground from the left hand was measured with a Megger BMM2000. The value was 3.3 MΩ with 100 V measurement voltage. Measurements were made only for one potential and capacitance value, because the results could be easily scaled to another level. Also, below certain voltage level discharges are difficult to achieve. 2.2 HBM ESD test In the HBM type of test the PWB and the test set-up were similar to the CDM ESD test, but the direction of current was opposite. The measuring person was insulated from ground and charged to 200 volts. The potential of the PWB was neutralised before a discharge. The charged person touched the needle by the tip of the hand tool and current waveform was recorded. Voltage increase on PWB’s surface was measured with TREK 541 electrostatic voltmeter and it was only 3–4 volts due to the high capacitance of the PWB. 2.3 Resistance of the hand tools Hand tools for the study were selected to have a wide range of parameters. Resistance from the handle of the hand tool to the tip was measured, and the order of the tools in Table 1 is

Figure 1 Sketch of the test set-up

© 2004 by Taylor & Francis Group, LLC

299

A B C D E F G

Description

Resistance (Ω)

Metal tweezers Snipe nose pliers with ESD handles Metal tweezers with carbon fibre tip Carbon fibre tweezers Snipe nose pliers with ESD handles Snipe nose pliers with ESD handles Snipe nose pliers with insulating handles

0 110 130 640 800 7,6 k 500 M

Table 1 Resistance of the tested hand tools

based on these measurements. The measurement method for the hand tool resistance was a modification of the method represented in [1]: Handles of the hand tool were compressed between jaws of a table vice. Conductive rubber sheet was placed between the vice jaw and the hand tool handles to ensure proper electrical connection, and a gold plated brass plate of 5×10×1 mm was placed between the jaws of the hand tool. The torque used to tighten the table vice was controlled. Then the resistance between the body of the table vice and the brass plate was measured with a Megger BMM2000. The lowest resistance (0 Ω) was for metallic tweezers and the highest for ordinary (non-ESD) snipe nose pliers (500 MΩ). 3. Results 3.1 CDM ESD test Discharge measurements were repeated five times for every hand tool. The hand tools were grounded and the PWB was charged to 100 volts before a discharge. An average of five discharges was calculated both for the peak current and for the charge (Figs. 2 and 3). Reproducibility of the results was good, but the charge calculated from the discharge current waveform was only a small part of the total charge stored in the capacitance of the PWB. In the measurements the main interest was in the initial peak of the discharge current, and with that focus, the measurement was not able to cover the entire duration of the discharge. The peak current is mainly determined by the capacitance of the PWB and the hand tool. Current transfer from a capacitor to another is very fast due to very small resistance and inductance. Remaining part of the energy stored in the PWB is transferred to ground through the person. This remaining part of the charge transfer is slow compared to charge transfer

Figure 2 Peak currents of the CDM ESD test for the hand tools

© 2004 by Taylor & Francis Group, LLC

300

Figure 3 Charge transfer in the CDM ESD test for the hand tools

between the PWB and the hand tool. Charge transfer through the human body is on a different time scale than the initial fast discharge, and it is not easily measured during the same measurement and with the same high frequency current probe as the initial peak. Thus, either the peak current or the charge transfer could be measured “correctly” by a single measurement. Figures 4 and 5 represent typical CDM ESD waveforms for tools B and D. 3.2 HBM ESD test The HBM ESD test for hand tools was quite similar to the CDM test. The main characteristics of the charge transfer are determined by the capacitance of the hand tool as in the CDM ESD test. In the HBM case the ratio of the capacitances and the direction of the current were opposite to the CDM case. The capacitance of the hand tools was very small compared to the capacitance of the four layer PWB. Therefore the voltage increase in the PWB was only 3–4 volts in spite of 200 V potential in a hand tool and the measuring person. The capacitance of a hand tool was calculated to be 4 pF with 4 V potential increase in the PWB. Figures 6

Figure 4 CDM ESD waveform for tool B

© 2004 by Taylor & Francis Group, LLC

Figure 5 CDM ESD waveform for tool D

301

Figure 6 Peak currents of the HBM ESD test for the hand tools

and 7 represent peak current and charge transfer values of HBM type of discharge. Figures 8 and 9 represent typical HBM ESD waveforms for tools B and D.

Figure 7 Charge transfer in the HBM ESD test for the hand tools

Figure 8 HBM ESD waveform for tool B

© 2004 by Taylor & Francis Group, LLC

Figure 9 HBM ESD waveform for tool D

302 4. Discussion The discharge current waveform in hand tool measurements can be divided into two parts. The first part is the fast initial charge transfer between the capacitances of a PWB and a hand tool, and the second, slower part, is the charge transfer where the human body is involved. With metallic tool tips the duration of the first current peak is in the order of 1 ns or less. The duration of the second part is 100–300 ns. In a single discharge only part of the total charge stored in the PWB capacitance is transferred. After the current transfer between the capacitances is completed, potential difference between the PWB and the hand tool is small and current may stop flowing. In some of the measured ESD waveforms this sudden stop of current flow was clearly visible. The potential of the PWB was also monitored in some of the CDM type of discharges with an electrostatic voltmeter, and the remaining potential of the PWB varied between 90 and 0 V. So in some cases charge was totally discharged from the PWB, but in some cases the discharge was only partial leading to a potential drop of only 10 V. In spite of that, measured charge did not vary much because only the fast initial discharge was measured. ESD tolerance values of electronic components are given in volts for different types of discharge models. If this voltage is known, it is possible to calculate peak current and charge threshold values for ESDS components [2]. By measuring peak currents and transferred charge values, in situations where PWB handling with a hand tool is necessary, it is possible to know how close we are to the ESD withstand level of the components and what is the risk of ESD damage. If the safety margin is very narrow, it is important to choose a correct hand tool. 5. Conclusions The only way to reduce peak current in the first contact discharge is to avoid metal to metal contact. Among the tested hand tools there were two tools with non-metallic parts at the tip. Both tweezers had low peak current values if compared to hand tools with metallic tips. Also the amount of charge transfer was smaller for tweezers with non-metallic tips. Ordinary pliers with the highest resistance from handles to the tip had the lowest charge transfer due to the minimum charge transfer between the human body and the hand tool. Acknowledgements The authors thank Heta Kojo at VTT for technical assistance. The work is partially supported by the Finnish National Technology Agency Tekes via the STAHA programme. References [1] [2]

International Electrotechnical Commission 1998 Technical Report 61340–5–2 Electrostatics— Protection of electronic devices from electrostatic phenomena—User guide Smallwood J and Paasi J 2003 Assessment of ESD threads to electronic components and ESD control requirements, Electrostatics 2003, p. 247–252.

© 2004 by Taylor & Francis Group, LLC

303

The role of capacitance in corona-electrode arrangements Charles G Nolfa, Nathaniel R Greeneb, Seth T Ashmanb, Michael A Catinob a XiPro Technologies LLC, Hatfield, PA 19440–0332 b Bloomsburg University, Hartline Science Center, Bloomsburg, PA 17815 Abstract. The surface charge density on a corona electrode is modified by adjustments to its capacitance with respect to the environment. The capacitance, electric flux per applied potential, is discussed in terms of concentric spheres and various cylinder and cylinder-plane electrode arrangements. Experimental data is presented to illustrate the influence of capacitance on corona onset conditions in the parallel wire-cylinder electrode configuration. Conductive shields were found to enhance the current transfer from an ionizer to a target surface in closely spaced electrode arrangements. The shield enhances ionization by increasing the capacitance and charge density on the corona electrode. At spacings larger than 50–100 mm, the corona emitter sees a free-space environment that is largely geometry independent.

1. Introduction The corona-electrode/counter-electrode geometry of typical commercial ionizers is typically complex and shaped for applications. Since the corona ionizers operate at voltages from 5,000 V to well above 50,000 V, protective shields are added for structural and safety reasons. The shields are often reported to enhance the uniformity and delivery of charge to surfaces. Support for this claim is neither obvious nor widely discussed in the research literature. More than 99% of ions generated by corona systems pass between the electrodes of closely spaced electrode assemblies, and consequently, do not participate in intended applications. The transfer of ions to gas improves with larger electrode spacing, where electric fields are weaker and exposure of the ions to the gas stream is greater. In many applications, including electrostatic painting, high-voltage transmission, and room ionization for ESD control, corona ionization occurs in “free space” and the location of the counter charge is unclear. Typically, the meaning or location of “ground” or “earth” is often neglected, avoided, or assigned to an arbitrarily positioned boundary for model closure. In this work the capacitance C of a corona emitter will be defined through Gauss’s law as the net electric flux per unit applied potential over its surface: C=⌽/V0 where is the net electric flux, and V0 is the potential difference between the electrodes. In the case of concentric spherical electrodes the capacitance tends to a constant value C∞=4πε0a, where the radius of the outer sphere is taken to infinity and the inner sphere has a radius a. As the radius of the distant outer sphere is changed, the flux, or electric field,

© 2004 by Taylor & Francis Group, LLC

304 about the sphere then remains essentially constant, yielding reduced sensitivity of corona processes about the smaller sphere to the geometric environment. There are known feedback mechanisms from photoionization at counter electrodes, but these generally involve more closely positioned counter electrodes than those for free-space electrodes of interest here. Aside from concentric spheres, electrode systems show dependence of the flux on the shape of the counter electrode. Generally, there is a strong geometric sensitivity for closely spaced electrodes, and this sensitivity becomes weaker with further electrode separation. Chow and Yovanovich [1] have shown the local geometric shape has relatively insignificant influence on the electric fields distant from the smaller electrode. Chakraborty, et al. [2], use moment methods to estimate capacitance and potential distributions for simple freespace electrodes. The capacitive environment of nearly “free-space” corona emitters is a second objective of this research. In this paper we explore the capacitance of two-dimensional electrode systems, and in particular the electric fields and fluxes about wires near and distant from various counter electrodes. Our preliminary studies have shown that local changes in electrode capacitance will intensify ionization from corona emitters [3]. The relatively simple measurement of capacitance is then found to be a powerful tool in ionizer development for exploring corona onset conditions. In a sense, the practical range for influence of the counter electrode is the overall goal of this work. 2. Analytical models Four electrode geometries are considered in the present work: coaxial cylinders, cylinderplane, parallel cylinders of equal radius, and parallel cylinders of unequal radius. Key variables are the radius of the first cylinder (a), the radius of the second cylinder (c), and the distance from the axis of the first cylinder to the nearest point on the counter electrode (d). The nomenclature was defined for the purpose of comparing the influences of the various counter-electrode arrangements on the first cylinder. Proceeding through the four electrode geometries opens the corona electrode to a more free space environment and increasing d illustrates the influence of taking the counter electrode away. The capacitances per unit length for these electrode sets have been calculated [4]. Coaxial cylinders

Cylinder-plane

Parallel cylinders of equal radius

Parallel cylinders of unequal radius

© 2004 by Taylor & Francis Group, LLC

305 Figure 1 plots C1, C2, and C3 for a 0.18 mm-diameter cylinder (a=0.09 mm) and its counter electrode as a function of d. As the counter electrode is moved more distant from the cylinder, the capacitance falls in each case—more rapidly at smaller distances. The capacitance is highest when the counter electrode surrounds the cylinder, takes on an intermediate value when the cylinder is parallel to a conductive plane, and is smallest for parallel cylinder systems. Increasing the size of the electrode (variable a) reduces the capacitance values, but preserves the trends discussed above. Figure 2 compares the capacitances C4 of a 5-mm-diameter cylinder as the radius of a second, parallel cylinder is increased, while keeping the distance d from the center of the first cylinder to the surface of the second cylinder constant. As the second cylinder’s radius grows, it subtends a greater solid angle as viewed by the wire, contributing to the wire’s capacitance. At very large radii, the second cylinder begins to play the role of a planar counter electrode, explaining the plateaus in the capacitance curves. Indeed, the expression for C4 approaches that of C2 in the limit of a large second-cylinder radius c.

Fig. 1. Capacitance of a 0.18-mm-diameter cylinder inside a larger cylinder, near a plane, or near an equal-diameter cylinder.

3. Experimental results The estimates of capacitance in Section 2.0 contain an assumption that all electrode capacitance comes from the defined counter electrode. It is clear from these calculations that the capacitance changes slowly at distances greater than about 50 mm from the “corona” electrode and that the capacitance is highest when the counter electrode geometry surrounds the wire. It is then not surprising that in room environments the capacitance of a typical corona emitter tends toward a constant value as the counter electrode is moved away. The corona emitter is small compared to the environment and bears the capacitance of an isolated sphere or point electrode. 3.1 Experimental arrangement A series of simplified experiments were performed to illustrate phenomena associated with cylinder-cylinder and cylinder-plane electrode configurations. The experimental arrangement in its most complete form is illustrated in Fig. 3. It is described in Ref. 3.

© 2004 by Taylor & Francis Group, LLC

306

Fig. 2. Capacitance of a 5-mm-diameter wire spaced 5, 10, and 20 mm from a second cylinder of variable radius.

Fig. 3. General arrangement of equipment.

3.2 Capacitance of cylinder parallel to a conductive plane Figure 4 compares the capacitance of a 4.75-mm-diameter brass cylinder and a 0.18-mmdiameter Ni-Cr wire (each 250 mm in length) and each oriented parallel to a grounded plane as a function of their separation d. The capacitances tend to nearly constant values at larger spacing.

© 2004 by Taylor & Francis Group, LLC

307 3.3 Capacitance of two cylinders above a conductive plane In the second series of experiments the 4.75-mm-diameter cylinder was placed 119.5 mm above the plane surface. A second cylinder of equal diameter was placed above and in contact with the first cylinder. The combined capacitance of these cylinders with respect to the plane was 5.55 pF. As the second cylinder was raised above the first (keeping its electrical connection to it), the capacitance of the cylinder pair increased towards a value near twice that for the cylinders in contact. The capacitances are additive and are those for two freespace cylinders.

Fig. 4. Capacitance of a 4.75-mm-diameter cylinder and a 0.18-mm-diameter wire above a flat metal surface.

3.4 Influence of capacitance on corona ionization The average electric field near a wire of radius a at potential V0 is proportional to its capacitance per unit length C, relative to its entire surroundings, For fine wires the actual distribution of this field over the surface of the wire does not strongly influence corona onset. It is the average intensity of the field, brought on by its capacitance and attendant surface charge that yields local gas breakdown to ionization. The experimental arrangement shown in Fig. 3 was used to explore the influence of wire capacitance on corona ionization. The wire was positioned 110 mm above the grounded surface, and a cylinder placed beneath it a distance d1=11.7 mm. When a second cylinder is positioned a distance d2 above the wire and moved towards it, the capacitance of the wire with respect to the two cylinders increases. In particular, the capacitance between the wire and the lower cylinder remains constant and the capacitance between the wire and the upper cylinder increases. Thus, as d2 decreases, with fixed voltage on the wire with respect to the grounded cylinders, the charge density on the wire increases. This occurs while the charge density on the lower cylinder remains constant and that on the upper cylinder increases. Under normal operation as an ionizer, the wire is in corona. The charge density on the wire then determines ionization with control and stabilization served by the counter electrodes and space charge. The counter-electrode cylinders in the above arrangement primarily serve as current-collectors. The increase in charge density brought on by the increased capacitance raises the electric field near the wire and, thus, the ionization current.

© 2004 by Taylor & Francis Group, LLC

308 From this simplified model, the function of the shield—the second cylinder—is not to project additional ions to the counter electrode, but rather to enhance current to both the shield and counter electrode. 3.5 Capacitance and corona onset voltage as predictors of current enhancement Let Il represent the wire current to the lower cylinder with the upper cylinder absent (d2/d1= ∞). When the upper cylinder is positioned symmetrically above the wire (d2/d1=1), the effect is to increase the lower cylinder current to Il’. From Fig. 5, this current enhancement is measured to be 80%.

Fig. 5. Current to cylindrical current collectors with d1=1.17 cm.

4. Conclusions Conductive shields are found to enhance the current transfer from an ionizer to a target surface in closely spaced electrode arrangements. The shield enhances ionization by increasing the capacitance and charge density on the corona electrode. At spacings larger than 50–100 mm, the corona emitter sees a free-space environment that is largely geometry independent. Here again, altering the capacitance in the emitter region can increase ionization. References [1] [2] [3] [4]

Chow Y L and Yovanovich M M 1982 J. Appl. Phys. 53, 8470–5 Chakraborty C, Poddar D R and Chakraborty A 1993 IEEE. Trans, on EM Compatibility 35, 98–102 Noll C G, Ashman S T, Catino M A and Greene N R 2001 Proc. Ann. Mtg. Electrostatics Society of America Page L and Adams N I 1960 Principles of Electricity: An Intermediate Text in Electricity and Magnetism (Princeton, D. van Nostrand Co.) 89–92

© 2004 by Taylor & Francis Group, LLC

309

Electrostatics and the Environment G S P Castle Department of Electrical and Computer Engineering, University of Western Ontario, London, ON, Canada, N6A 5B9 Abstract. Electrostatic phenomena play a very important role in the environment both in nature and in various industrial processes used to alleviate pollution. This paper is intended to review some of these electrostatic phenomena and applications with emphasis on three particular areas; lightning, electrostatic precipitation, and plastics separation/recycling. Lightning represents the oldest manifestation of electrostatics and the most spectacular demonstration of the power of electrical forces in nature. Surprisingly, in spite of the fact that it has been studied for many years, differences of opinion still exist as to its cause. The current understanding of the mechanisms and theories are reviewed. Electrostatic precipitation is well known being the first significant industrial application of electrostatics. It is used extensively to remove particulate contaminants from air. Current challenges exist in attempting to extend the process to also deal with gaseous pollutants. These integrated approaches are highlighted including the combined use of precipitators with scrubbers, pulsed discharge plasma reactors and other advanced oxidation techniques. Electrostatic techniques have been preferentially used for separating mixtures of conductive and nonconductive materials for recycling. The current challenge is to try to extend these principles to separation of dissimilar nonconductive materials such as plastics. One successful approach to this problem has been developed at the University of Western Ontario. The system is described and some results of recent industrial separations are reported.

1. Introduction The topic of electrostatics and the environment covers such a broad field that it is impossible to do justice to all aspects in the brief space allotted to this discussion. For example, electrostatic phenomena play very important roles both in nature and in various industrial processes used to alleviate air, water and soil pollution. Within these general categories there are a great many specific examples that can be identified ranging from the controversial, such as the effect of air ions on human health, through to the unusual, such as the use of electric fields for strengthening unstable soils by extracting water from clays. Rather than attempting a comprehensive review of all of the areas encompassed by this topic the intention of this discussion is to highlight three particular areas of general interest and special importance to the environment; lightning, electrostatic precipitation and plastics separation/recycling. 2. Lightning Lightning represents the oldest manifestation of electrostatics and the most spectacular demonstration of the power of electrical forces in nature. It can even be argued that it is possibly the most important single factor in our environment as it has been suggested that energy released through early lightning strikes was instrumental in the creation of life on earth. In pioneering experiments reported in 1953 [1], Nobel laureate Harold Urey’s student, S.L.Miller demonstrated that when a gas mixture simulating what is believed to be the constituents of the early earth’s atmosphere (H2O, H2, CH4, and NH3) was exposed to a continuous electric discharge for one week, trace amounts of amino acids were found to be

© 2004 by Taylor & Francis Group, LLC

310 produced. These materials are the building blocks of organic life. Clearly this possibility should be the starting point in any discussion of the interconnection between electrostatics and the environment. However as engineers and physicists we are not normally too comfortable in dwelling on this subject as it is more within the domain of biochemists and perhaps even theologians and philosophers. We are more concerned with the nature of lightning itself. Who amongst us has not been both fascinated and awed by the sheer splendour and power of a full-blown thunder, lightning storm? Since the earliest of times lightning has been observed and its mechanism puzzled over. 2.1 Importance of Lightning to the Environment There are multiple reasons why lightning is so important to us in our natural environment. For a start it can be deadly. Approximately 1000 people die each year from lightning strikes. (Interestingly, Brazil the world’s largest tropical country accounts for over 10% of these fatalities [2]). Furthermore lightning represents the largest single cause of power and telecommunication failures. However, like all things in nature, it is not only something to be feared. For example although the majority of forest fires are primarily initiated by lightning and cause much tragedy and economic damage each year, biologists understand that such destruction is a natural and necessary part of the health and regeneration of forests [3]. From the point of view of electrostatics however the significance of lightning is very clear, it maintains the electrical equilibrium of the earth. Under normal atmospheric conditions, the earth is negatively charged and experiences an average “fair weather” electric field of approximately 100 V/m, in the downwards direction. This is accompanied by a positive ion current flow of 10-12 A/m2 (approximately 1800 A world wide). [4] On its own, this would result in a neutralization of the negative charge on the earth within minutes and elimination of the electric field that we know exists in steady state. Lightning is the mechanism that stabilizes this phenomenon by replenishing the negative charge on the earth. As I speak, throughout the world on average, over 1000 thunderstorms are occurring resulting in 100 lightning strikes per second. These localized electrical discharges allow positive charge to flow from the earth into the atmosphere resulting in potential equilibrium. What is the cause of these discharges? 2.2 Mechanism of Lightning Scientific study began in earnest with the “enlightening” experiments of Benjamin Franklin in the late 1700’s and has subsequently been studied in detail for many years. However surprisingly, in spite of this long period of study even today we do not fully understand it. Differences of opinion still exist as to the causes of this most common natural phenomenon. What we do know is that in a normal thunderstorm, clouds develop in which the lower part of the cloud is polarized with negative charge and the upper part with positive charge. On first thought this does not present a situation that should result in an electric discharge since the earth is also negatively charged. However, the lower part of a thundercloud will normally have a much larger negative potential than the earth underneath resulting in a strong negative electric field that attracts negative electrons towards the earth. This process has been well studied and results in the well known “step leader”, the sporadic, irregular passage of fast electrons moving in 50 m spurts, stopping for approximately 50 microseconds, then continuing in an erratic path to the nearest point on the ground. Once this leader connects with ground a complete conducting path exists between the cloud and earth. The leading electrons are the first to flow to ground leaving behind an electron depleted region consisting of positive ions that then attracts more electrons from further up the leader. This © 2004 by Taylor & Francis Group, LLC

311 progresses rapidly upwards along the conducting channel so that all of the negative charges adjacent to the point of origin in the cloud flow rapidly to ground through this “return stroke”. This takes place very rapidly (typically the return stroke lasts 0.1ms and travels at a speed of 38× 106m/s) and very energetically (typically consisting of 20 C and a peak current of 10 kA). The remaining negative charge in the cloud can redistribute very quickly and travel down the remains of this ionized path (the “dark leader”) creating another return stroke. This repeats several to tens of times resulting in the liberation of sound and light energy which we know as thunder and lightning. Of course this is a very simplistic and incomplete description of this complex phenomenon since it is known that approximately 10% of lightning strikes are actually “positive” lightning and much lightning occurs from cloud-to-cloud or even internal to a single cloud. Also great controversy exists about the nature (and according to some even the existence) of “ball” lightning [5]. However perhaps the most surprising thing about our state of knowledge is the incomplete understanding about how the charge separation in the cloud comes about in the first place. Many theories and experiments have attempted to answer this and the two most accepted generally are based upon either particle interactions or inductive processes.[6] In the former it is believed that interactions among ice crystal, graupel and liquid water lead to charge separation and then segregation in the cloud due to differing masses of the respective materials. In the induction process it is believed that an ambient electric field causes polarization of particles that then separate. Clearly the latter situation requires a pre-existing electric field, which then leads to the conclusion that probably both processes are important. Nature however keeps some secrets well and it would appear that much work remains to be done to come to a definitive answer to this question. 3. Electrostatic Precipitation Moving from nature and the environment to the effects of industrialization and the environment, it is well known that the first significant industrial application of electrostatics was electrostatic precipitation (ESP). [7] It is currently used extensively both in industry and domestic applications to control air pollution by removing particulate contaminants from air. Although we now recognize ESP as one of the most important technologies for air pollution control, it is ironic to note that the first commercial installation which took place in California almost 100 years ago was primarily motivated by economic considerations related to the recovery of valuable catalysts rather than protection of the environment. 3.1 Importance of ESP to the Environment Currently throughout the world there are many thousands of installations of ESP’s cleaning effluents from major industrial sources such as coal-fired power plants, smelting plants etc. In addition, two stage electrostatic precipitators are widely used to remove small particulates from recycled air in industrial and domestic applications. Although the principles of ESP technology are well-understood, continual improvements in terms of collection efficiency, reliability and cost effectiveness have taken place, primarily in the last 35 years or so. Driven by increasingly stringent environmental regulations major advances have occurred with the development of techniques to improve the collection for high resisitivity materials and for very fine particles in the size range less than 10µm. These changes have included improvements in; the mechanical design of the precipitators, the aerodynamics of the air flow, the use of conditioning agents to reduce particle resistivity and induce agglomeration, the use of new power sources that allow intermittent and pulse energization to improve performance and save energy, and the use of “on line” measurement and control to maintain optimum performance under varying conditions.[8] © 2004 by Taylor & Francis Group, LLC

312 Current challenges exist in attempting to extend the process to also deal with gaseous pollutants such as SOx and NOx. At this time the only feasible way to remove such gases is to use an ESP to remove the particles followed by absorption in a liquid scrubber. Other noxious gases such as volatile organic compounds (VOC) and dioxins are currently removed by incineration. However much ongoing effort is being expended in trying to induce chemical reactions using such techniques as non-thermal discharge plasmas, electron beam reactors, and advanced oxidation techniques[9]. Unfortunately, economic and reliability issues are significant barriers to adoption and much more work will be required before these become viable alternatives. 4. Plastics Separation/Recycling As a final example of the importance of electrostatics to the environment, I would like to highlight a relatively new technology, the separation of plastic waste. It is well known that electrostatic techniques have been preferentially used for many years for separating mixtures of conductive and nonconductive materials for recycling. [10] However the current challenge is to try to extend these principles to the separation of dissimilar nonconductive materials such as plastics. This process has the major advantage of using dry materials, unlike for example alternative methods such as flotation. The importance of the need for this type of recycling can be highlighted by the fact that in North America paper and paperboard have a recycling rate of over 40%. These materials are of course renewable resources. Yet it is energy efficient and beneficial to the environment to recycle them. On the other hand, plastics, which are produced from nonrenewable and rapidly depleting petrochemical resources, are recycled at an overall rate of only 5%.[11] The rest is disposed of in landfill or in some other such wasteful or nonproductive manner. It is true that certain types of plastics such as used in plastic bottles are recycled at a higher rate. (Post consumer bottle recycling was approximately 22% in the USA in 2001.) However bottles represent an easily separated and relatively clean product. More generally, post consumer plastics are mixed with many contaminants and contain several types of plastic. The separation problem is complicated by the mutual incompatibility of the most widely used commercial plastics which include Polyvinyl Chloride (PVC), Polyethylene Terephalate (PET), high and low density Polyethylene (HDPE & LDPE), Polypropylene (PP), and Polystyrene (PS). All these plastics have special individual properties and as a result find uses in many different products. Separation purities must be very high in order to enable the material to be reused in place of virgin material. Thus there is a real demand for a process that can selectively segregate mixtures of two or more types of plastic efficiently and economically. 4.1 Principle of Separation At the IOP Electrostatics 1991 meeting we presented a paper describing the feasibility of separating certain types of commercial grade plastics using contact charging in a fluidized bed and dropping them through a horizontally oriented electric field. [12] A disadvantage of this process is that fluidization requires significant mechanical energy. Since then this work has been extended to produce a more practical and energy efficient separator involving simply a rotating inclined cylindrical drum. In this system plastic particles, typically consisting of two or three different constituents chopped or ground to a size of the order of 5 mm, are fed through the rotating drum and undergo contact charging with each other and to a lesser extent with the inside wall of the cylinder. This equipment evolved from a laboratory scale separation unit that consisted of a 10cm diameter grounded metal charging drum, 1m in length, and having one end closed. This was mounted above a 3m high separation tower in which the © 2004 by Taylor & Francis Group, LLC

313 high voltage electrodes were mounted. The electrodes were connected to variable +/- 60kV power supplies. Depending upon the voltages applied, either symmetric or asymmetric horizontal electric fields of up to 300kV/m were possible. A series of nine shielded collection trays having their long axis perpendicular to the electric field were placed at the base of the tower. In this way both the mass and charge of each collected fraction were measured. The separation tower acts as a macroscopic mass spectrometer where the greater the charge to mass ratio (Q/M) the further the particles are deflected from straight vertical fall. An extensive series of laboratory scale batch experiments were carried out in which a weighed quantity (typically 200g) of two or three component mixed plastic material was rotated (typically at 5–20 rpm) for a fixed time (typically 2 minutes) with the drum slightly inclined upwards. The cylinder was then tilted slightly downwards (about 10 degrees) and the particles gradually fell out of the drum into the tower and were collected in the trays. The Q/M and composition of the components was determined either by using different coloured constituent materials or by x-ray spectrography. The quality of separation was measured by determining the sample purity (mass of desired component in collected sample/ total mass in collected sample) and the efficiency was measured by the sample recovery (mass of desired component in collected sample/total mass of initial material). By appropriate adjustment of the electric field strength and inclusion of samples from various numbers of the collection trays adjacent to either the positive, negative or grounded field electrodes it was possible to determine different levels of material purity and recovery. Detailed results were presented elsewhere [13]. Selected examples are shown in Table 1.

Table 1. Separation results for selected mixtures of virgin plastic mixtures

4.2 Industrial Applications Plas-Sep Ltd., London, ON, Canada, has subsequently extended the excellent results found in the laboratory tests into industrial scale applications. A unit module capable of handling a continuous feed of one tonne/hour uses a charging drum 0.5m in diameter and 2m long. The fall distance is 4m. Two adjustable separators that collect 3 components, positively charged, negatively charged and middlings are used in place of the nine collection trays. The polarity

© 2004 by Taylor & Francis Group, LLC

314 of the charge that develops depends upon the materials and various versions of effective ‘Triboelectric Series” have been published for different plastics. However a number of factors normally complicate the prediction of chargeability. For example, such things as surface printing dyes, colorants and contaminants from the previous use for the plastic can dramatically affect the charging polarity and magnitude. However in many practical cases for given combinations of materials although not predictable a priori the charging properties are consistent. Two factors affect this. The first is that the fact that the plastics need to be chopped thus exposing a considerable amount of virgin plastic surface. The second is the need to ensure that the relative humidity (RH) of the air in which the plastic is treated is controlled. Normally a RH<50% is satisfactory. Successful and economically attractive separations have been achieved in post-industrial waste applications for two-component mixtures in the automobile industry (tail light and bumper assemblies) and the electrical industry (chopped wire insulation.) Even more significant are recent results that have successfully separated three component plastic mixtures from dashboard scrap.[14] The technology is also suitable for separating many types of post-consumer waste plastic material although the necessity to pre-treat and clean the material makes the economics less attractive. 5. Conclusions Clearly it is impossible to do full justice to a topic as broad as electrostatics and the environment in this brief discussion. However it is hoped that the three specific examples presented here demonstrate the fundamental importance of this topic in nature, pollution control and conservation of our resources. Much essential work remains to be done both in understanding and applying electrostatic phenomena in the environment. 6. References 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14)

S.L.Miller, “A Production of Amino Acids Under Possible Primitive Earth Conditions”, Science, Vol 117, No 3046, May 15,1953, p528 O.Pinto, “Lightning in Brazil Sets Shocking Record”, www.cnn.com/2002/TECH/science/1 1/ 05/brazil.lightning.reut/index.html R.W.Gorte, “Forest Fires and Forest Health”, CRS Report for Congress, 95–511 ENR, National Council for Science and Environment, Washington, DC, July 14, 1995 R.P.Feynman, R.B.Leighton and M.Sands, “Chapter 9:Electricity in the Atmosphere”, The Feynman Lectures on Physics, Volume II, Addison-Wesley Publishing Co., Reading, Mass, sixth printing, 1977, pp9–1–11 D.J.Turner, “Ball Lightning and Other Meteorological Phenomena”, Physics Reports 293, Elsevier, 1998, pp 1–60 K.Miller et al, “Modeling and Observations of Thundercloud Electrification and Lightning”, Atmospheric Research 58, Elsevier, 2001, pp89–115 H.J.White, “Centenary of Fredrick Gardner Cottrell”, J. Electrostatics, 4, 1977/78, pp1–34 K.R.Parker, “Applied Electrostatic Precipitation”, Chapman and Hall, London, 1997 H.H.Kim et al, “Low Temperature NOX Reduction Processes Using Combined Systems of Pulsed Corona Discharge and Catalysts”, Journal of Physics D: Applied Physics, 34, 2001, pp604–137 H.Higashiyama and K.Asano, “Recent Progress in Electrostatic Separation Technology”, Particulate Science and Technology, Vol 16, #1, 1998, pp77–90 R&D Magazine, Cahners Business Information, Highlands Ranch, CO, Sept 2002, p54 I.I.Inculet and G.S.P.Castle, “Triboelectrification of Commercial Plastics In Air”, Institute of Physics Conference Series, #118, Oxford, 1991, pp217–222 I.I.Inculet, G.S.P.castle and J.D.Brown, “Electrostatic Separation of Plastics for Recycling”, Particulate Science and Technology, Vol 16, #1, 1998, pp91–100 J.D.Brown, “Electrostatic Plastics Separation”, Recycling Technology Newsletter, Natural Resources Canada, Vol 5, #1, 2000, pp1–2

© 2004 by Taylor & Francis Group, LLC

315

Measurement methods in electrostatics applications: review and trends W D Greason Department of Electrical and Computer Engineering, Faculty of Engineering, University of Western Ontario, London, Ontario, N6A 5B9, Canada Abstract. A review of evolving measurement techniques in applied electrostatics is presented for the period from the 1999 Institute of Physics Conference on Electrostatics, held in Cambridge, to the present. The main sources referenced include the Journal of Electrostatics, the Electrostatic Processes Committee contributions to the IEEE Transactions on Industry Applications and the Conference Proceedings of the Electrical Overstress/Electrostatic Discharge Symposia. Measurements in electrostatics applications can vary from the basic, employing standard instruments with a relatively small number of recordings to the complex, involving many specialized instruments used to obtain data intensive records of events. Several areas of interest include: non-invasive measurements employing signature analyses principles which do not disturb the phenomena under study, investigation of microgap phenomena involved in magnetoresistive recording heads and MEMS technology, data acquisition and analyses in sensor intensive applications recording high frequency transient events associated with electrostatic discharges and the study of charge decay on insulators including the characterization of spatial charge distribution profiles. Enhanced measurement methodologies are clearly evolving; our educational institutions are encouraged to include the subject of measurement of electrostatic phenomena in courses on electronic instrumentation and measurement.

1. Introduction A review of measurements related to the following topics is presented: charge, discharges, medium characterization, ESPs and reactors, probes and transducers, biotechnology, MEMS and space exploration. 2. Charge The measurement of surface potential decay on dielectric materials has been described [1– 4]; a space-charge measuring system based on a laser-induced pressure pulse (LIPP) method along with thermally stimulated discharge current (TSDC) measurements can be applied to analyze charge distribution in thin films [5,6]. Space charge, electric field and potential in multi-layered dielectric films can be studied from interfacial charges measured using the step-electroacoustical (SEA) technique [7,8]. The measurement of bipolar charge in powders [9] and on aerosols [10,11] has been described. Measurement methodologies for the charge carrier extraction from gas streams [12], charge distributions of particles in gas-solids pipe flow [13] and streamer inception in liquids using a differential charge measurement system have been presented [14].

© 2004 by Taylor & Francis Group, LLC

316 Experimental systems to study corotron charging behaviour in electrophotographic processes have been summarized [15,16]. Apparatus to investigate various aspects of triboelectrification have been described [17–19]. A reading unit with the ability of noncontact charge detection was used to demonstrate the feasibility of an electret floppy disk for digital information storage [20]. 3. Discharges 3.1 General Electrostatic discharges leading to explosions are a frequent cause of industrial accidents. Experimental approaches to study ignition tests with brush discharges [21], the ignition of methane-air mixtures by multiple capacitor discharges [22] and measurements of incendivity of electrostatic discharges from textiles used in personal protective clothing [23] have been reported. The following are examples of measurement based studies on the general topic of discharges: bipolar dc corona discharge from a floating filamentary metal particle [24]; electrical discharge occurring from a space charge cloud formed with charged particles [25]; characterisation of electrostatic discharges from insulating surfaces [26]; NOx production in spark and corona discharges [27]; correlation between the optical signatures and current wave forms of long sparks with applications in lightning research [28]; methodology to study the resistance of spark discharges [29]. 3.2 Corona Experimental apparatus and measurement procedures have been described for dc corona discharges [30–32], corona discharges and a water environment [33–39], pulsed corona discharges [40–43], corona discharge in the pin-to-plane geometry [44–47] and ozone generation in ac corona discharge [48]. 3.3 ESD ESD continues to be a major reliability issue in electronic devices and systems. Test methods and methodologies continue to evolve [49–51]. Measurement of the indirect effect of ESD offers a non invasive means to characterize the event; detection of the EMI due the radiated electric and magnetic fields has been studied [52–60] as well as the use of a calorimeter to detect the optical signal emitted by a spark discharge [61,62]. Other measurement related topics include: broadband measurement of ESD risetimes [63]; fields on horizontal coupling planes excited by direct ESD to the vertical coupling plane [64]; interferometric temperature mapping during ESD stress of protection devices [65]; surface resistivity and charge decay measurements of materials [66]. 3.3.1 GMR Heads Giant magnetoresistive heads used in the disk drive industry are very susceptible to damage from low level ESD transients. Procedures for the testing and measurement of ESD susceptibility have been described and documented as follows: tribocharging and electrical breakdown at the head-disk interface [67]; ESD sensitivity of GMR heads at variable pulse length [68]; measurement of current transients [69]; considerations for ESD standards for measuring and testing of MR heads [70]; electrostatic voltmeter and fieldmeter measurements on GMR heads [71]; field-induced charged device model testing of MR heads [72].

© 2004 by Taylor & Francis Group, LLC

317 3.3.2 TIP Transmission line pulsing has become established as a measurement method to characterize the performance of ESD protection devices. Measurement set ups for various applications have been reported: TLP ESD testing of GMR heads [73]; TLP testing of low voltage triggering SCR devices [74]; analysis of the switching behaviour of ESD protection transistors [75]; breakdown and latent damage of ultra thin gate oxides [76]; verification of ESD protection device response [77]; TLP of integrated structures [78]. The design of TLP test sets, calibration and evolving standards have also been described [79–80] 4. Medium Characterization 4.1 Dielectrics Measurement methods to study the properties of dielectric materials have been reported and include: interdigital dielectrometry for the non-destructive evaluation of material properties [81]; high-temperature conduction of plasma-treated films [82]; non-linear conductivity spectra of antistatic polymer films [83]; influence of electron-beam irradiation on electrical parameters of dielectric materials [84]. 4.2 Powders and Particles Experimental apparatus and measurement methods have been presented for powder coatings, fluidized beds, the charging of particles and powders and phenomena related to the charge on powders and particles. For powder coatings, work reported includes: adhesion measurements for coatings using drop test rig and virtual oscilloscope [85]; measurement of thickness and adhesive properties [86]; electroseparation and efficiency of deposition [87]; powder flow control system with capacitance sensor [88]. Two papers on fluidized beds [89,90] provide detail on the instrumentation and measurement methods employed. Aspects of the charging processes include: unipolar charging of aerosol particles in alternating electric field [91]; change in charging characteristics of polymer powder by plasma treatment [92]; characterization of chargeability of biological particulates by triboelectrification [93]. Work related to the charge on particles and powders include: real-time particle size and charge distribution analysis [94]; evaluation of charge retention properties of powder paints by thermally stimulated current spectroscopy [95]; free air beam in an electric field for determination of charge on powders [96]; trajectories of charged aerosol particles near a spherical collector [97]; electrostatic measurements on lactose-glucose mixtures [98]; static charge elimination on pellets in a silo using a nozzle type eliminator [99]. 4.3 Sprays Recent publications include the description of measurement set ups for experiments on the spraying of liquids [100–102] and ceramic suspensions [103–105]. Some topics of interest include: development of electrostatic pressure-swirl nozzle for agricultural applications [106]; electrostatic application of pollen sprays [107]; electrohydrodynamic spraying characteristics of glycerol solutions in vacuum [108]. 4.4 Liquids Recent papers which discuss measurement principles related to liquids can be broadly classified as follows: charge generation, liquid characterization, specific measurements.

© 2004 by Taylor & Francis Group, LLC

318 Charge generation has been described for impeller mixing of used transformer oil [109] and the internal electrification of diesel oil injectors [110]. Liquid characterization processes include: deformation and break-up of aqueous drops of dielectric liquids in high electric fields [111]; electrohydrodynamic surface waves of thin oil films generated by needleplate barrier discharges under reduced gas pressure [112]; surface tension reduction of liquid by applied electric field using vibrating jet method [113]. Specific measurement procedures reported include: power consumption measurements for ac and pulsed dc for electrostatic coalescence of water-oil emulsions [114]; measurements of surface conductivity in dielectric liquid [115]; Kerr electro-optic field mapping measurements [116]; the use of FET models for analysing electro-osmosis [117]. Papers of special interest include liquid dielectrophoresis on the microscale [118] and electrostatic actuation of liquid droplets for microreactor applications [119]. 5. Large Scale Operations 5.1 ESP Experimental arrangements and instrumentation for measurement have been described for pulsed and dc electrostatic precipitator systems [120–123]. Other subjects include: smoke precipitation by charged water aerosol [124]; effect of particle diameter and corona electrode geometry on the particle migration velocity in electrostatic precipitators [125]; pseudoelectret filter for micrometer-sized particles in exhaust gases [126]; submicron charged dust particle interception by charged drops [127]. 5.2 Reactors Non-thermal plasma processing is one of the most promising technologies to remove toxic gas contaminants in air; several publications describe the experimental set ups, instrumentation and measurement methods employed in the experiments [128–142]. Other topics involving the description of measurements include: measurement of hydroxyl radicals in an atmospheric pressure discharge plasma using laser-induced fluorescence [143] improvement of NOx removal efficiency using atomization of fine droplets in corona discharge [144]; plasma-assisted chemical process for NOx control [145]; removal of NF3 from semiconductor-process flue gases by tandem packed-bed plasma and adsorbent hybrid systems [146]; reduction of NOx from natural gas combustion flue gases by corona discharge radical injection techniques [147]; indoor air cleaning using a pulsed discharge plasma [148]. 6. Probes/Transducers The design and characterization of PVDF transducers [149,150] and electrostatic flow probes [151–155] has been documented. Other probes developed for specific measurement applications include: vibrating-wire transducers for electrostatic measurements [156]; floating double probe characteristic utilized to measure electron temperature [157]; the use of Tassicker’s measurement formula in the boundary biased-probe method [158]; visualization of the distribution of electron traps on polymer surfaces using a scanning micro laser probe [160].

© 2004 by Taylor & Francis Group, LLC

319 7. Biotechnology The use of electrostatics in the field of biotechnology is an emerging topic of interest. Measurement methodologies have been reported in the following work: selective detection of viable bacteria using dielectrophoretic impedance measurement method [161]; effect of culture temperature on high-voltage pulse sterilization of escherichia coli [162]; integrated planar concentric ring dielectrophoretic (DEP) levitator [163]; electrical sterilization of escherichia coli by electrostatic atomization [164]; enhancing effect of electret on transdermal drug delivery [165]; molecular surgery of DNA based on electrostatic micromanipulation [166]; inactivating microorganisms using a pulsed electric field continuous treatment system [167]; quantitative analysis of DNA orientation in stationary ac fields using fluorescence anisotrophy [168]. 8. MEMS Microelectromechanical systems (MEMS) technology is in its development stage and will present measurement challenges in the future. Two applications have been described: a MEMS field mill made from two surface micro-machined polysilicon thin film layers deposited over a silicon substrate [169]; electrostatic micromirrors for subaperturing in an adaptive optics system [170]. 9. Space Exploration NASA has an interest in dust particle charging on Mars; the design and development of instruments for in situ measurements presents a great challenge. The design of an electrometer to evaluate the electrostatic nature of the Martian soil and atmosphere has been presented [171]; an experimental apparatus to evaluate the triboelectrification of a Martian soil simulant has been reported. Future work will include the development of particle charge spectrometers to measure the size and charge on individual dust particles and an instrument to measure electrostatic fields generated by rover interactions with the Martian surface. 10. Trends and Needs Instrumentation and measurement are essential components of research based programs. As new electrostatic based technologies evolve, there is a need for enhanced measurement methodologies. Non-invasive measurements which do not disturb the phenomenon under study are of interest since instrumentation loading effects are eliminated. The study of microgap phenomena involved in magnetoresistive recording heads and MEMS technology present measurement challenges. Wide bandwidth probes for electrostatic discharge measurements are needed to characterize current waveforms and charge transfer. There is a need for integrated measurement devices for sensor intensive applications. The use of advanced instrumentation and measurement principles in electrostatics research presents a challenge to educators. More than ever, one must not only know how to measure a physical parameter but also be able to interpret the results to know what is being measured. Our educational institutions are encouraged to include the subject of measurement of electrostatic phenomena in courses on electronic instrumentation and measurement.

© 2004 by Taylor & Francis Group, LLC

320 References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44]

Sahli S, Bellel A, Ziari Z, Kahlouche A and Segui Y 2003 J. Electrostatics 57 169–181 Chubb J 2002 J. Electrostatics 54 233–244 Chubb J 2000 IEEE Trans. Incl Appl. 36 1515–1522 Chubb J 1999 Proc. IOP Conf. Ser. No. 163 329–333 Takashima K and Oda T 2000 IEEE Trans. Incl Appl. 36 76–81 Takashima K and Oda T 1999 IEEE Trans. Incl Appl 35 1192–1197 Motyl E 2001 J. Electrostatics 51–52 522–529 Motyl E 2001 J. Electrostatics 51–52 530–537 Zhao H, Castle G S P and Inculet 11 2002 J. Electrostatics 55 261–278 Kulon J, Hrabar S, Machowski Wand Balachandran W 2001 lEEETrans. Incl Appl. 37472– 479 Kulon J and Balachandran W 2001 J. Electrostatics 51–52 552–557 Noll C G, Miller W and Bracikowski C 2002 J. Electrostatics 54 283–291 Matsusaka S, Umemoto H, Nishitani M and Masuda H 2002 J. Electrostatics 55 81–96 Dumitrescu L, Lesaint O, Bonifaci N, Denat A and Notingher P 2001 J. Electrostatics 53 135– 146 Feng J Q, Morehouse P W and Facci J S 2001 IEEE Trans. Incl Appl 37 1651–1657 Hays D A 2001 J. Electrostatics 51–52 57–63 Greason W D, Kucerovsky Z and Zaharia C 2002 J. Electrostatics 56 385–398 Ohara K, Nakamura I and Kinoshita M 2001 J. Electrostatics 51–52 351–358 Yu Z and Watson K 2001 J. Electrostatics 51–52 313–318 Amjadi H and Franz C J.Electrostatics 50 265–277 Schwenzfeuer K and Glor M 2001 J. Electrostatics 51–52 402–408 Ptasinski L and Zeglen T 2001 J. Electrostatics 51–52 395–401 Kathirgamanathan P, Toohey M J, Haase J, Holdstock P, Laperre J and Schmeer-Lioe G 2000 J. Electrostatics 49 51–70 Asano K, Hishinuma R and Yatsuzuka K 2002 IEEE Trans. Incl Appl. 38 57–63 Higashiyama Y, Sugimoto T and Takahashi M 2001 J. Electrostatics 51–52 381–386 Davidson J L, Williams T J, Bailey A G and Hearn G L 2001 J. Electrostatics 51–52 374–380 Rehbein N and Cooray V 2001 J. Electrostatics 51–52 333–339 Gomes C and Cooray V 1998 J. Electrostatics 43 267–274 Greason W D 1999 IEEE Trans. Ind. App. 35 359–365 Noll C G 2002 J. Electrostatics 54 271–282 Noll C G 2002 J Electrostatics 54 245–270 Asano K, Yatsuzuka K and Yamaki T 2000 IEEE Trans. Ind. Appl. 36 87–92 Suarasan I, Ghizdavu L, Ghizdavu I, Budu S and Dascalescu L 2002 J. Electrostatics 54 207– 214 Higashiyama Y, Yanase S and Sugimoto T 2002 J. Electrostatics 55 351–360 Yamamoto M, Nishioka M and Sadakata M 2002 J. Electrostatics 56 173–187 Sugimoto T, Asano K and Higashiyama Y 2001 J. Electrostatics 53 25–38 Sugimoto T and Higashiyama Y 2001 J. Electrostatics 53 209–219 Fujii T, Aoki Y, Yoshioka N and Rea M 2001 J. Electrostatics 51–52 8–14 Moon J, Kim J and Lee D 1998 IEEE Trans. Ind. Appl. 34 1212–1217 Ono R and Oda T 2002 J. Electrostatics 55 333–342 Yan K, van Heesch E J M, Pemen A J M, Huijbrechts P A H J, van Gompel F M, van Leuken H and Matyas Z 2002 IEEE Trans. Ind. Appl. 38 866–872 Yan P, Fujiwara M, Zhou Y and Ishida M 2001 J. Electrostatics 51–52 266–271 Kirkpatrick M, Finney W C and Locke B R 2000 IEEE Trans. Ind. Appl. 36 500–509 Kanazawa S, Ito T, Shuto Y, Ohkubo T, Nomoto Y and Mizeraczyk J 2002 J. Electrostatics 55 343–350

© 2004 by Taylor & Francis Group, LLC

321 [45] Kucerovsky Z, Greason W D and Ieta A 2001 J. Electrostatics 50 147–157 [46] Dascalescu L, Samuila A, Rafiroiu D, Iuga A and Morar R 1999 IEEE Trans. Ind. Appl. 35 543–548 [47] Takahashi N, Goldman A, Goldman M and Rault J 2000 J. Electrostatics 50 49–63 [48] Feng R, Castle G S P and Jayaram S 1998 IEEE Trans. Ind. Appl. 34 563–570 [49] Chaine M, Verhaege K, A very L, Kelly M, Gieser H, Bock K, Henry L G, Meuse T, Brodbeck T and Barth J 1998 EOS/ESD Symposium Proc. EOS-20 301–310. [50] Carey R E and DeChiaro L F 1998 EOS/ESD Symposium Proc. EOS-20 40–53 [51] Lin D, Pommerenke D, Barth J, Henry L G, Hyatt H, Hopkins M, Senko G and Smith D 1998 EOS/ESD Symposium Proc. EOS-20 29–39 [52] Rudack A C, Pendley M and Levit L 2002 J. Electrostatics 54 95–104 [53] Ptasinski L, Zeglen T and Gajda J 1999 Proc. IOP Conf. Ser. No. 163 225–228 [54] Takai T, Kaneko M and Honda M 1998 J. Electrostatics 42 305–320 [55] Frei S and Pommerenke D 1998 J. Electrostatics 44 191–203 [56] Lin D L, DeChiaro L F and Jon M 1998 J. Electrostatics 44 159–175 [57] Bernier J, Croft G and Lowther R 1998 J. Electrostatics 44 149–157 [58] Jacksen N, Tan W and Boehm D 1998 EOS/ESD Symposium Proc. EOS-20 233–244 [59] Munoz J L, Tan J, Adriano C, Roldan E and Sadie J 2000 EOS/ESD Symposium Proc. EOS22 60–64 [60] Jacksen N, Nelsen L, Boehm D and Odom T 2000 EOS/ESD Symposium Proc. EOS-22 193– 201 [61] Kucerovsky Z, Greason W D and Flatley M W 2002 IEEE Trans. Ind. Appl. 38 1162–1167 [62] Kucerovsky Z, Greason W D and Flatley M W 2002 IEEE Trans. Ind. Appl. 38 77–84 [63] Bonisch S, Pommerenke D and Kalkner W 2002 J. Electrostatics 56 363–383 [64] Frei S and Pommerenke D 1998 J. Electrostatics 44 177–190 [65] Furbock C, Pogany D, Litzenberger M, Gornik E, Seliger N, Gobner H, Muller-Lynch T, Stecher M and Werner W 2000 J. Electrostatics 49 195–213 [66] Paasi J, Nurmi S, Vuorinen R, Strengell S and Maijala P 2001 J Electrostatics 51–52 429–434 [67] Himle J and Wallash A 2002 EOS/ESD Symposium Proc. EOS-24 142–146 [68] Guarisco D and Li M L 2000 EOS/ESD Symposium Proc. EOS-22 322–326 [69] Ogle W and Moore C 2000 EOS/ESD Symposium Proc. EOS-22 217–223 [70] Henry L G and Wallash A 2000 EOS/ESD Symposium Proc. EOS-22 337–342 [71] Pritchard D 2000 EOS/ESD Symposium Proc. EOS-22 499–504 [72] Wallash A J 1996 EOS/ESD Symposium Proc. EOS-18 8–13 [73] Wallash A 2002 J. Electrostatics 56 331–339 [74] Guilhaume A, Galy P, Chante J P, Foucher B, Bardy S and Blanc F 2002 J. Electrostatics 56 281–294 [75] Wolf H, Gieser H and Wilkening W 2000 J. Electrostatics 49 111–127 [76] Wu J, Juliano P and Rosenbaum E 2000 EOS/ESD Symposium Proc. EOS-22 287–295 [77] Hyatt H, Harris J, Alanzo A and Bellew P 2000 EOS/ESD Symposium Proc. EOS-22 111–120 [78] Gieser H and Haunschild M 7996 EOS/ESD Symposium Proc. EOS-18 85–94 [79] Chu C Y and Worley E R 1998 EOS/ESD Symposium Proc. EOS-20 311–319 [80] Barth J, Verhaege K, Henry L G and Richner J 2000 EOS/ESD Symposium Proc. EOS-22 85–96 [81] Mamishev A V, Takahashi A R, Du Y, Lesieutre B C and Zahn M 2002 J. Electrostatics 56 465–492 [82] Takanashi T and Ikezaki K 2002 IEEE Trans. Ind. Appl. 38 846–851 [83] Tajitsu Y 2002 J. Electrostatics 55 299–310 [84] Vila F and Sessler G M 2001 J. Electrostatics 51–52 146–152 [85] Dastoori K and Makin B 2001 J. Electrostatics 51–52 509–514 [86] Dastoori K, Makin B and Telford J 2001 J. Electrostatics 51–52 545–551

© 2004 by Taylor & Francis Group, LLC

322 [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129]

Barmuta P and Cywinski K 2001 J. Electrostatics 51–52 239–244 Yanagida K, Morita T and Takeuchi M 2000 J. Electrostatics 49 1–13 Alissa Park A, Bi H T, Grace J R and Chen A 2002 J. Electrostatics 55 135–158 Yao L, Bi H T and Park A 2002 J. Electrostatics 56 183–197 Lackowski M 2001 J. Electrostatics 51–52 225–231 Murata Y and Aradachi T 2001 J. Electrostatics 51–52 97–104 Banerjee S and Law S E 1998 IEEE Trans. Incl. Appl. 34 1201–1205 Mazumder M K, Grable N, Tang Y, O’Connor S and Sims R A 1999 Proc. IOP Conf. Ser. No. 163 335–340 Ogiwara T and Ikezaki K 1999 Proc. IOP Conf. Ser. No. 163 325–328 Epping R.H and Kuettner A 2002 J. Electrostatics 55 279–288 Jaworek A, Adamiak K, Krupa A and Castle G S P 2001 J. Electrostatics 51–52 603–609 Murtomaa M and Laine E 2000 J Electrostatics 48 155–162 Kodama T, Suzuki T, Nishimura K, Yagi S and Watano S 2002 J. Electrostatics 55 289–297 Huneiti Z and Balachandran W 2001 J. Electrostatics 51–52 558–564 Davies K and Gerken S 2000 J. Electrostatics 49 139–150 Huneiti Z A, Balachandran W and Machowski W W 1998 IEEE Trans. Ind. Appl. 34 pp. 279– 285 Miao P, Balachandran W and Xiao P 2002 IEEE Trans. Ind. Appl. 38 50–56 Miao P, Balachandran W and Wang J L 2001 J. Electrostatics 51–52 43–49 Balachandran W, Miao P and Xiao P 2001 J. Electrostatics 50 249–263 Laryea G L and No S 2003 J. Electrostatics 57 129–142 Law S E, Wetzstein H Y, Banerjee S and Eisikowitch D 2000 IEEE Trans. Ind. Appl. 36 998– 1009 Ku B K and Kim S S 2003 J. Electrostatics 57 109–128 Wagner J P and Clavijo F R 2000 J. Electrostatics 50 31–47 Romat H and Badri A 2001 J. Electrostatics 51–52 481–487 Eow J S, Ghadiri M and Sharif A 2001 J. Electrostatics 51–52 463–469 Lee T S and Chang J S 2001 J. Electrostatics 51–52 565–570 Sato M, Kudo N and Saito M 1998 IEEE Trans. Ind. Appl. 34 294–300 Lee C M, Sams G W and Wagner J P 2001 J. Electrostatics 53 1–24 Matsubara Y 1999 J. Electrostatics 46 125–130 Gung T J, Ustundag A and Zahn M 1999 J. Electrostatics 46 79–89 Watanabe S, Yamada Y, Hobo Y, Hayashi N, Sumi T, Dykes D and Touchard G 2001 /. Electrostatics 51–52 455–462 Jones T B 2001 J. Electrostatics 51–52 290–299 Washizu M 1998 IEEE Trans. Ind. Appl. 34 732–737 Elayyan H S B, Bouziane A and Waters R T 2002 /. Electrostatics 56 219–234 Zukeran A, Looy P C, Chakrabarti A, Berezin A A, Jayaram S, Cross J D, Ito T and Chang J S 1999 IEEE Trans. Ind. Appl. 35 1184–1191 Liang X, Looy P C, Jayaram S, Berezin A A, Mozes M S and Chang J S 2002 IEEE Trans. Ind. Appl. 38 69–76 Mok Y S, Lee H W and Hyun Y J 2001 J. Electrostatics 53 195–208 Balachandran W, Krupa A, Machowski W and Jaworek A 2001 J. Electrostatics 51–52 193– 199 Jedrusik M, Gajewski J B and Swierczok A J 2001 J. Electrostatics 51–52 245–251 Inculet 11, Castle G S P, Slanina M and Duca M 2002 IEEE Trans. Ind. Appl. 38 64–68 Jaworek A, Krupa A and Adamiak K 1998 IEEE Trans. Ind. Appl. 34 985–991 Oda T J. 2003 Electrostatics 57 293–311 Yamamoto T, Okubo M, Nagaoka T and Hayakawa K 2002 IEEE Trans. Ind. Appl. 38 1168– 1173

© 2004 by Taylor & Francis Group, LLC

323 [130] Kuroki T, Takahashi M, Okubo M and Yamamoto T 2002 IEEE Trans. Ind. Appl. 38 1204– 1209 [131] Okubo M, Tanioka G, Kuroki T and Yamamoto T 2002 IEEE Trans. Ind. Appl. 38 1196–1203 [132] Kim H H, Prieto G, Takashima K, Katsura S and Mizuno A 2002 J. Electrostatics 55 25–41 [133] Song Y, Kim S, Choi K and Yamamoto T 2002 J. Electrostatics 55 189–201 [134] Moon J D, Lee G T and Geum S T 2000 J. Electrostatics 50 1–15 [135] Ogata A, Yamanouchi K, Mizuno K, Kushiyama S and Yamamoto T 1999 IEEE Trans. Ind. Appl. 35 1289–1295 [136] Ogata A, Shintani N, Mizuno K, Kushiyama S and Yamamoto T 1999 IEEE Trans. Ind. Appl. 35 753–759 [137] Shimizu K and Oda T 1999 IEEE Trans. Ind. Appl. 35 1311–1317 [138] Moon J D, Lee G T and Chung S H 1999 IEEE Trans. Ind. Appl. 35 1198–1204 [139] Okumoto M, Su Z, Katsura S and Mizuno A 1999 IEEE Trans. Ind. Appl. 35 1205–1210 [140] Yamamoto T and Jang B W L 1999 IEEE Trans. Ind. Appl. 35 736–742 [141] Kohno H, Berezin A A, Chang J S, Tamura M, Yamamoto T, Shibuya A and Honda S 1998 IEEE Trans. Ind. Appl. 34 953–966 [142] Okumoto M, Rajanikanth B S, Katsura S and Mizuno A 1998 IEEE Trans. Ind. Appl. 34 940– 944 [143] Ono R and Oda T 2000 IEEE Trans. Ind. Appl. 36 82–86 [144] Ohneda H, Harano A, Sadakata M and Takarada T 2002 J. Electrostatics 55 321–332 [145] Yamamoto T, Yang C L, Beltran M R and Kravets Z 2000 IEEE Trans. Ind. Appl. 36 923– 927 [146] Chang J S, Kostov K G, Urashima K, Yamamoto T, Okayasu Y, Kato T, Iwaizumi T and Yoshimura K 2000 IEEE Trans. Ind. Appl. 36 1251–1259 [147] Urashima K, Chang J S, Park J Y, Lee D C, Chakrabarti A and Ito T 1998 IEEE Trans. Ind. Appl. 34 934–939 [148] Mizuno A, Kisanuki Y, Noguchi M, Katsura S, Lee S W, Hong Y K, Shin S Y and Kang J H 1999 IEEE Trans. Ind. Appl. 35 1284–1288 [149] Janiczek T 2001 J. Electrostatics 51–52 167–172 [150] Mazurek B, Janiczek T and Chmielowiec J 2001 J. Electrostatics 51–52 76–81 [151] Gajewski J B 2000 J. Electrostatics 48 279–294 [152] Gajewski J B 1999 J. Electrostatics 46 271–284 [153] Gajewski J B 1999 J. Electrostatics 48 49–64 [154] Gajewski J B 1999 J. Electrostatics 45 255–264 [155] Kacprzyk R and Gajewski J B 2001 J. Electrostatics 51–52 124–130 [156] Struminsky V I 2002 J. Electrostatics 54 301–310 [157] Liu C,Wang J, Yu K, Eliasson B, Xia Q, Xue Band Zhang Y 2002 J. Electrostatics 54 149–158 [158] Jayaram S, Xu X Y and Cross J D 2000 IEEE Trans. Ind. Appl. 36 920–922 [159] Shi X, Yumoto M and Sakai T 1999 IEEE Trans. Ind. Appl. 35 549–553 [160] Murata Y and Kubota K 1999 Proc. IOP Conf. Ser. No. 163 341–344 [161] Suehiro J, Hamada R, Noutomi D, Shutou M and Kara M 2003 J. Electrostatics 57 157–168 [162] Ohshima T, Okuyama K and Sato M 2002 J. Electrostatics 55 227–235 [163] Qian L, Scott M, Kaler K V I S and Paul R 2002 J. Electrostatics 55 65–79 [164] Lee H 2001 J. Electrostatics 51–52 71–75 [165] Cui L, Jiang J, Zhang L, Song C, Zhao W and Lin J 2001 J. Electrostatics 51–52 153–158 [166] Yamamoto T, Kurosawa O, Kabata H, Shimamoto N and Washizu M 2000 IEEE Trans. Ind. Appl. 36 1010–1017 [167] Qin B, Barbosa-Canovas G V, Swanson B G, Pedrow P D and Olsen R G 1998 IEEE Trans. Ind. Appl. 34 43–50 [168] Suzuki S, Yamanashi T, Tazawa S, Kurosawa O and Washizu M 1998 IEEE Trans. Ind. Appl. 34 75–83

© 2004 by Taylor & Francis Group, LLC

324 [169] Horenstein M N and Stone P R 2001 J. Electrostatics 51–52 515–521 [170] Horenstein M N, Pappas S, Fishov A and Bifano T G 2002 J. Electrostatics 54 321–332 [171] Buchler M, Cheng L, Orient O, Thelen M, Gompf R, Bayliss J and Rauwerdink J 1999 Proc. IOP Conf. Ser. No. 163 189–196 [172] Gross F B, Grek S B, Calle C I and Lee R U 2001 J. Electrostatics 53 257–266

© 2004 by Taylor & Francis Group, LLC

325

Pulsed arc discharges for water treatment and disinfection H Z Zastawny1, H Romat1, N Karpel Leitner2, J S Chang3 Laboratoire d’Etudes Aerodynamiques, Electrohydrodynamic group, Bd Pierre et Marie Curie, Téléport 2, BP 30179, 86962 Futuroscope-Chasseneuil, France 2 Laboratoire de Chimie de 1’Eau et de FEnvironnement, 40 av. du Recteur Pineau 86000 Poitiers, France 3 McMaster University, Hamilton, Ontario, Canada

1

Abstract. Within the framework of a collaboration between our laboratory and McMaster University, a physical process based on the properties of pulsed arc discharges within water has been investigated for the treatment and the disinfection of water. The arcs are produced in a cylindrical tank between two rod-to-rod electrodes. In this article we first present the experimental device and then electrical results obtained (current and potential measured from one of the electrodes) for different kinds of water, for varying applied potentials and for several electrode gaps in the water. We also give the results that we obtained on the pressure waveforms from piezoelectric pickups placed at two different distances from the arc zone. Finally we analyse all these results in order to better understand the propagation of the shock waves produced by the pulsed arc system.

1. Introduction Most of the actual processes used for the treatment or the disinfection of water and which are based on chemistry properties generate by-products which are increasingly regulated by the European Union. Alternative processes have been studied during the last decade by a certain number of authors in order to find a compromise between the efficiency and the environmental impact. Most of the research workers [1–3] have studied the electroacoustical efficiency of an electrical discharge in water. J.S.Chang et al. [4] have particularly investigated the pulsed arc discharges in water and showed that an arc could generate a strong shock wave, cavitation bubbles as well as strong ultraviolet radiation and OH radicals. They tested this process on zebra mussels empirically and successfully. This paper focuses on the characterization of the electrical pulsed arc parameters and on their influence on the generated pressure. Arcs are created in a cylindrical transparent pilot between two electrodes, one of them being connected to a capacitor via a high voltage spark gap switch. When this external switch allows current to circulate, all the electric charge of the capacitor is commuted to a rod-to-rod system in the water, which produces a plasma in the water and a compulsive pressure pulse [5].

© 2004 by Taylor & Francis Group, LLC

326 2. Experimental arrangements The experimental device, which is shown in figure 1, comprises a cylindrical tank, a pulsed arc generator and the measurement system. The tank is made of Plexiglas and two rod-torod type electrodes are placed on its revolution axis. The production of pulsed arcs between these electrodes requires an electrical system which is composed of a capacitor (50 µF), its charge circuit and a high voltage spark gap switch. The latter is composed of two rod-to-rod type electrodes which are separated by air. The level of applied potential is fixed by the gap between these electrodes. When the potential reaches the disruptive value for the air, an arc is created in the air and a current can circulate through the air and after within water. The voltage and the current are measured from the input electrode with respectively a voltage divider type sensor and a coil type sensor, both linked to a Tektronix oscilloscope which can record the signals.

Figure 1—Schema of the experimental apparatus

3. Electrical characterization of the discharge for different media In a first series of experiments we tested different media such as tap water and solutions of ferric or permanganate ions. For different electrode and spark gap values, we measured the applied voltage and the discharge current from which we deduced the electrical power. 3.1. Influence of the discharge on the media First we did tests on a violet solution of permanganate ions (KMnO4) of concentration 0.1 g.L-1. After one hour of discharges at the rate of one arc every 3 seconds, the final concentration measured by colorimetry techniques was 0.05 g.L-1. Next we did the same experiment with a solution of ferric ions (FeCl3) whose initial pH was 3. After one hour of treatment the pH was 5 and the colour of the solution corresponded to a decrease of 1/3 of the concentration of ferric ions. These two rudimentary experiments show that arc discharges generate reactions within liquid containing reducing agents (decrease of the concentration of ferric ions and permanganate ions). The ferric ion (Fe3+) reduction involves the production of OH and Os radicals which permit the elimination of some pollutants like pesticides [6].

© 2004 by Taylor & Francis Group, LLC

327 3.2. Electrical parameters Figure 2 shows a typical recording of the voltage, the current and the electrical power obtained whatever the media in which arcs are produced, or the electrode gap and the applied voltage. When the capacitor discharges through the spark gap switch, the potential of the input electrode rises abruptly and then decreases regularly until breakdown, which corresponds to the circulation of a weak conduction current in the water (a few hundreds of amperes is not high compared to the thousands of amperes of maximum current during the discharge!). When the breakdown occurs, the potential decreases rapidly to zero while the current assumes the shape of a peak.

Figure 2—Typical voltage, current and electrical power recording

3.3. Parameters influencing on the discharge The value of the conduction current before breakdown is always higher for the ferric solution than in the other cases; it is 5% of the value of the current peak observed after the breakdown within the ferric solution but 1% of it in the case of the permanganate solution. The solution of ferric ions is more conductive than the permanganate or water. Figure 3 (a) gives the evolution of the voltage stage time before the breakdown as a function of the electrode gap between the electrodes in water when Vmax (potential at the beginning of the stage) is equal to 4 kV. For a fixed electrode gap we can see that the lowest time necessary before breakdown is for the ferric solution, which is likely to be due to the conductivity of the liquid. In fact the duration of the stage before the arc discharge depends on the power injected into the water by the conduction current. For a fixed potential, if the conductivity of the liquid is high, its resistance is low and the power injected in the liquid (V2/R) is important. So the system does not need as much time in this case as in the other ones (with lower conductivity) to initiate the arc. We also notice that whatever the discharge media, the more the electrode gap increases, the longer the voltage stage lasts. When the gap increases, the resistance of the water increases too, and for a fixed potential the power injected decreases. The time necessary is then higher when the distance between the electrodes increases.

© 2004 by Taylor & Francis Group, LLC

328 Figure 3(b) shows the results obtained for the voltage stage time as a function of Vmax when the electrode gap is fixed at 1mm. We see that the voltage stage time decreases when Vmax grows. For the same reason when Vmax increases the power injected increases too and the time necessary to initiate the process of discharge diminishes (the arc starts when the energy injected in the water is sufficient to initiate the process).

Figure 3—Voltage stage time as a function of the electrode gap (a) and as a function of Vmax (b)

Figure 4—Electrical energy released per discharge as a function of the electrode gap (a) and as a function of Vmax (b)

Figure 4(a) gives the mean energy released during the discharge as a function of the gap between the electrodes in the water for an applied potential Vmax of 4 kV. The mean electrical energy per discharge has been calculated from the time integral of the electrical power visualised on the oscilloscope. Whatever the media, this energy decreases when the electrode gap grows. As shown in figure 5(b) for which the gap is fixed to 1 mm, it increases when the applied potential Vmax increases.

© 2004 by Taylor & Francis Group, LLC

329 The power released during the discharge is proportional to the applied potential and to the inverse of the length of the plasma (resistance of the medium). In both graphs of figure 5 we can see that indeed the energy (the mean power multiplied by the duration of the discharge which is constant in any case: 50 µs) decreases when the length of the plasma (the gap) increases and increases when the applied potential increases. We also notice in these graphs that the energy released is dependant of the solution. The reason for this is not yet understood. 4. Pressure outline for underwater arcs In order to understand the propagating pressure wave phenomena, pressure measurements were carried out with arcs produced within water for a fixed electrode gap and for varying spark gap values. As shown in figure 1, the piezoelectric pressure sensor were placed on the lateral circular wall of the tank, 100 mm from the arc zone, or on the top of the cylinder, 50 mm from the arc zone.

Figure 5—Typical pressure wave, potential, current recording

A typical visualization of the pressure wave phenomena on the oscilloscope is shown in figure 5. The presence of the pressure stage before the breakdown is due to electrical interferences and must not be considered here. The pressure waveform starts 34 µs after the discharge, as shown in the right-hand graph of figure 5, which corresponds roughly to the theoretical time of the sound wave propagation within the water [7]. Its evolution shows a sequence of peaks with an observable time period. This pressure waveform may represent a spherical shock waveform which includes many reflections on the tank walls. Nevertheless we must notice on the right-hand side graph of figure 5 that the shape of the first peak of the pressure is very similar to the one of the current or of the power. As shown in the figure 6, when the electrode gap is set to 2 mm and the potential Vmax is equal to 4 kV, the first pressure peak we can observe after the discharge is about 10 bars when the pressure is measured 100 mm from the arc zone and about 23 bars when it is measured 50 mm from the arc zone. Although the distance between the sensor and the arc may not be the only factor influencing the pressure value and although the sensor does not face the arc similarly in both positions the first pressure peak always increases with the potential Vmax. The more the

© 2004 by Taylor & Francis Group, LLC

330

Figure 6—Pressure peak as a function of Vma

arc injects energy into the liquid, the higher the first peak is. Further experiments in which the electrode gap and the sensor position will be varied are planned in the future. 5. Conclusion Pulsed arc discharge generation involves reactions within polluted water: the presence of certain oxidants induces the production of useful radicals for the eradication of some pollutants. Arc production within a liquid is influenced by the properties of the media, the applied voltage and the electrode gap. The combination of these various parameters acts on the setting up of the discharge: the resistance between the electrodes determined by the electrode gap and the specific conductivity of the environment influences the passage of the conduction current and induces the presence of a slow decreasing voltage stage. Pressure wave propagation is one of the mechanisms generated by pulsed arcs. In our experimental case, these pressure waves may have the shape of a spherical shock wave which reflects on the pilot walls. In future work, a numerical simulation of this wave type propagation will be compared to the experimental profile. References [1] [2] [3] [4] [5] [6] [7]

Roi N A and Frolov D P 1958 Soviet Physics: Doklady 118–121 Gavrilov G N et al. 1977 Soviet Physics: Technical Physics 22 868–870 Okun’ I Z 1971 Soviet Physics: Technical Physics 16 216–226 Chang J S, Looy P C and Urashima K 2000 IEEE Conference on Electrical Insulation and Dieletric Phenomena 10 105–108 Martin E A 1960 J. Appl. Phys. 31 n°2 225–267 Foret C 2002 DEA Université de Poitiers LCEE Mortimer B J P, Skews B W, Felthun L T 1998 Shock Waves 8 63–69

© 2004 by Taylor & Francis Group, LLC

331

Electrostatic Spray Application of Decontaminant Agents onto the Human Body as a Bioterrorism Countermeasure: Process Development and Evaluation S Edward Law1, Steven C Cooper2 and Mark A Harrison3 Department of Biological and Agricultural Engineering, University of Georgia, Athens, GA 30602–4435, USA 2 Electrostatic Spraying Systems, Inc., 62 Morrison St., Watkinsville, GA 30677, USA 3 Department of Food Science and Technology, University of Georgia, Athens, GA 30602–7160, USA

1

Abstract. Electrostatic-induction charging of conductive liquids has been developed as an effective method for rapidly applying decontaminant sprays (e.g., antitoxins, disinfectants, sanitizers, etc.) and medicinal sprays onto the human body as a protective countermeasure against chemical and biological agents contaminating cutaneous surfaces. Aqueous-based sprays having typically -16 mC/kg charge-to-mass droplets of ~30 µm volume-median diameter were electrostatically deposited by action of spacecharge and image-charge electric fields and quantified fluorometrically for 52 body locations on a test mannequin as well as on a similar size human subject. Averaged over all body sites, charged spray deposited significantly (α< 0.0001) greater areal density of liquid than did identically dispensed uncharged sprays, providing ~ 1.8-fold increase. The electrodeposition benefit ranged from ∼1.0- to 3.8-fold for specific body sites, and charged sprays exhibited greater uniformity of deposition over the body (e.g., the coefficient of variability was 29% and 40%, respectively, for charged and uncharged spray applications). Additionally, preliminary microbiological assessment, via deactivation of an innocuous bacterium (Pseudomonas fluorescens) inoculated onto body target sites prior to spraying a broad-spectrum quaternary ammonium decontaminant, showed that ~2.38 log10 reduction in colony-forming units was provided by the electrostatically-enhanced countermeasure vs. only ~0.66 logic reduction for similarly applied uncharged decontaminant, indicating ~50-fold greater bactericidal efficacy for the electrostatic method.

1. Introduction Recent world events have prompted reassessment and refinement of civilian-defense strategies and capabilities for whole populations [1]. In the case of airborne dispersal, there exists potential need for an effective countermeasure and associated apparatus capable of rapidly decontaminating chemically and biologically active agents deposited onto cutaneous surfaces of the human body. System requirements dictate thorough spatial coverage of the skin with decontaminant yet minimal waste and disposal of the material. Electrostatic spray application of finely atomized droplets has been shown to meet similar deposition requirements in numerous applications ranging from finish coatings onto manufactured goods to chemical and biological pest-control agents onto agricultural crops [2]. This latter

© 2004 by Taylor & Francis Group, LLC

332 system [3] has been developed having spray-charging attributes especially compatible with the engineering and safety constraints necessary for applications onto humans, namely: lowvoltage/low-current/low-capacitance electrostatic-induction spray-charging nozzles operating at 1 kV dc or less providing 10–15 mC/kg spray charge-to-mass values; chargedspray generation by pneumatic atomization of conductive liquids (e.g., 10–4–10 S/m) into 20–30 µm droplets having residual aerodynamic energy to penetrate the electrified droplets into Faraday-shielded regions (e.g., underarm skin) with subsequent space-charge-field deposition; and liquid flowrates in the 1–2 mL/s range. This paper reports the design and experimental evaluation of a portable, microprocessor controlled, walk-through booth facilitating electrostatic application of decontaminant (e.g., antitoxins, disinfectants, sanitizers, etc.) and medicinal sprays onto the human body. Design operational conditions are 90 persons/h, 100 mL/person charged spray dispensed, and 20 mL/person liquid-waste disposal. Skin-deposition assessments include fluorescent tracer documentation of spatial distribution and areal density of spray mass, as well as assessment of microbiological efficacy of the electrostatic process via deactivation of an innocuous microorganism applied to the skin prior to spraying an appropriate decontamination agent. 2. Experimental Analysis The overall experimental facility for electrostatically applying sprays onto test subjects is depicted by Fig. 1. It consisted of a dielectric (high-density polyethylene) spray chamber having a grounded mesh-steel floor over a dielectric sump and a grounded aluminum footplate on which the subject stood. An assembly of three induction-charging spray nozzles oscillated at approximately 1.13 rad/s in the vertical plane to provide a 14 s sequence often 75° head-to-foot spray sweeps over the standing body facing the nozzles spaced 1 m away. During a short interlude the subject was rotated 180° and the spray sequence was repeated. Following a 30 s quiescent period for residual spray deposition and/or evacuation, the subject exited the chamber. Not shown is an attached 1.05 m wide×0.65 m long×2.08 m high process-support room housing ancillary equipment for spray-liquid metering, pneumatic and induction-voltage inputs to the charging nozzles, mechanical oscillation, and a microprocessor controller for properlysequencing all operations. Detailed descriptions of additional apparatus, materials, procedures, and statistical design of the experiment follow. 2.1. Apparatus and Materials 2.1.1. Spray-charging. Spray liquid of 0.1 S/m conductivity was prepared by combining 1.5 g of fluorescent tracer particles (DayGlo Corp.—Blaze Orange GT15N), 0.1 gNaCl, and 1.0 mL of Triton X-100 non-ionic surfactant into 1000 mL of deionized water. The liquid was pneumatically atomized (186 kPa) at 1.19 mL/s per nozzle into ~30 µm volume-median diameter spray and inductively charged to approximately –16 mC/kg by an embeddedelectrode spray-charging nozzle (Electrostatic Spraying Systems, Inc.— MaxCharge™ model) operating at 1.1 kV dc. Fig. 1. Electrostatic spray chamber. Similar nozzles typically exhibit a 10–15 uC/ kg’V linear spray-charging response vs. input voltage [4]. © 2004 by Taylor & Francis Group, LLC

333 For the microbiological assessment, the spray liquid consisted of a 0.78% v/v tap-water solution of the quaternary ammonium disinfectant Coverage Plus® (Merck and Co., St. Louis, MO; EPA Reg. No. 6836–78–1043) containing Di-n-alkyl dimethyl ammonium chloride and N-alkyl dimethyl benzyl ammonium chloride. Atomization and spray-charging performance were similar to the tracer spray liquid above. 2.1.2. Target systems. A commercial mannequin was utilized to simulate a human subject for most spray applications. The polymeric surface was rendered conductive by spray coatings of a water-based EMF-shielding paint (Staticveil®, Less EMF Inc., Ghent, NY); surface resistivity on test cells measured 2 ohms (i.e., ohm/sq) and typical resistance to earth was <1.5 kΩ from the head and other body locations. Figure 2 shows the 1.85 m tall male mannequin, erect standing, with arms modified for 20° outward orientation vis-àvis the body’s central axis. Circular targets denote some of the 52 deposition sites evaluated. To facilitate replicated spray applications and to eliminate background fluorescence attributable to the mannequin’s surface, stainless steel flat washers (17.5 mm OD×5.16 mm ID×1.60 mm thick) were affixed to the surface using flat-head stainless steel screws; an identical backing washer was installed behind each target washer to preclude surface contamination from the mannequin. When installed, the effective target area of each test washer was 2.07 cm2. Abbreviated sites (viz., 11) were utilized for the microbiological assessment and the human subject. For the former, target washers were autoclaved for sterilization, inoculated with the bacteria, and installed with a sterile backing washer onto the mannequin using sterile forceps. For the latter, double-sided paper tape, with a copper-tape grounding jumper, attached a target washer at each of the designated human-body sites; negligible background fluorescence from the tape’s adhesive to the target backside was confirmed. Fig. 2. Test mannequin. 2.2. Experimental Procedure 2.2.1. Fluorometric analysis. Spray tracer deposited onto the targets was reclaimed into 10 mL of wash liquid (0.1% Triton X-100 surfactant in deionized water) by 20 min agitation in 15 mL wash vessels [5]. Tracer deposition and its areal concentration (ng/cm2) were subsequently quantified using a digital fluorometer (Turner model 450–005) exciting the tracer by quartz-halogen irradiation (λ<440 nm) and measuring emission (λ>535 nm) to include the tracer’s 602 nm dominant fluorescence. Calibration standards established a linear (R2=0.999) fluorometer response. Four liquid samples from the wash cell for each respective sprayed target were fluorometrically analyzed and fluorometric mean values converted to (Spray-Liquid Deposition Density, nL/cm2)=3.08 (fluorometer reading)+33.55.

© 2004 by Taylor & Francis Group, LLC

334 2.2.2. Microbiological assessment. The ubiquitous, nonsporeforming, Gram negative, aerobic bacterium Pseudomonas fluorescens was selected for this assessment. A culture was grown in tryptic soy broth at 28° C for 20–22 h before inoculating 10 uL onto the front surface of each sterile test washer and drying in a laminar flow hood. Sterile procedures were observed for all handling of the targets including attachment and detachment from the mannequin. After spraying and detachment, each washer was placed into a 50 mL sterile centrifuge tube, 10 mL of sterile 0.1% peptone buffer added, and the tube shaken vigorously for 60 s before a portion of liquid from each was spiral plated onto plate-count agar plates. The plates were incubated at 28° C for 48 h before colony-forming units (CFU) of bacteria were enumerated. 2.2.3. Experimental design. Using a randomized complete block statistical design, the two spray treatments of the (52 body sites)×(2 charge conditions) were completed in a single day; four day-replications were made. Spray-charge condition was charged or uncharged. Experimental data were analyzed by SAS General Linear Models procedure using standard analysis of variance (ANOVA) to identify statistically significant differences in mean values of deposition density achieved by charged vs. uncharged application treatments onto individual target sites as well as onto sites averaged over the body both regionally and wholly. Coefficient of variability values were computed for estimates of deposition uniformity for charged vs. uncharged spray treatments. 3. Results and Discussion 3.1. Deposition Analysis Table 1 summarizes the mean values (n=4) for areal concentration (nL/cm2) of spray liquid deposited onto specified target sites for the two spray-charge treatments. Accompanying ANOVA results indicate which of the 52 treatment means significantly differ from one another. Figures 3 and 4 illustrate, respectively, the trends in areal deposition density for the uncharged vs. charged spray-application methods for the mannequin and the human subject.

Table 1. Mean values of spray-liquid deposition density onto designated mannequin and human body sites for uncharged and charged spray applications, (nL/cm2). © 2004 by Taylor & Francis Group, LLC

335 3.1.1. Mannequin deposition. The increases in spray deposition attributable to electrostatic-force effects are presented in Fig. 3 as ratios of charged deposition divided by uncharged deposition at each target site on the mannequin. These increases ranged from 0.98-fold for sites electrostatically shielded from the oncoming spray stream (e.g., site 36) to a maximum of 3.82-fold for difficult sites parallel to the air-velocity vector (e.g., site 37) for which electrostatic forces prove most beneficial in redirecting the spray droplets for deposition. Averaged over all 52 target sites, charged spray deposited Fig. 3. Electrostatic enhancement of deposition. significantly ( α < 0.0001) greater areal density of liquid (2159 nL/cm2) than did uncharged spray (1211 nL/cm2) for a 1.78-fold overall electrodeposition benefit. In addition to increasing the whole-body deposition, electrostatic effects improved the uniformity of surface coverage by spray; charged-spray applications exhibited a significantly (α=0.0567) lower coefficient of variability of 29% as compared with 40% for uncharged spray. 3.1.2. Zonal deposition. In addition to whole-body analysis, mannequin target sites were grouped into the following zones where detailed comparisons showed electrostatic spray deposition within all zones to be significantly (␣ < 0.01) improved in both quantity and uniformity: 1) Head/neck (sites #1–7)—1511 nL/cm2, cv 19% ch. vs. 764 nL/cm2, cv. 36% unch.; 2) Torso (#8–15, 18, 19, 21, 29–32)—2224 nL/cm2, cv 24% ch. vs. 1247 nL/cm2, cv 30% unch.; 3) Appendages including arms/hands (#16, 17, 20, 22–28, 33, 34)—1878 nL/ cm2, cv 25% ch. vs. 1057 nL/cm2, cv 26% unch. and legs/feet (#35–52)—2549 nL/cm2, cv 23% ch. vs. 1490 nL/cm2, cv 38% unch. 3.1.3. Human deposition. Figure 4 graphically compares areal concentrations of sprayliquid deposited at the 11 human-body sites by the uncharged and charged tracer sprays. Averaged over all sites, whole-body charged-spray deposition (1261 nL/cm2) exceeded uncharged-spray deposition (592 nL/cm2) by 2.13-fold. 3.2. Microbiological Assessment Plate-count results from the unreplicated applications of decontaminant spray indicated approximately a 47-fold greater reduction in bacteria CFU achieved by charged vs. uncharged treatments when averaged over all target sites; individual charged-spray sites exhibited typically only 0.1–9.7%) (median ~1.0%) surviving bacterial counts as compared with counts for corresponding uncharged-spray sites. Overall bactericidal efficacy, as expressed by log10 reductions in CFU from the typical 8 log10 value of the inoculated but unsprayed control targets, was approximately 2.38 log10 and 0.66 log10, respectively, for charged and uncharged applications of the 0.78% v/v aqueous spray of broad-spectrum disinfectant.

© 2004 by Taylor & Francis Group, LLC

336

Fig. 4. Charged and uncharged spray deposition achieved on human body.

4. Conclusion Incorporation of electrostatic forces of attraction significantly increased the overall spray deposition by ~2-fold onto cutaneous surfaces of the human body and onto a similar test mannequin. For individual target sites, the electrostatic deposition benefit ranged from ~1.2to 4.3-fold for the human body and ~1.0- to 3.8-fold over the 52 mannequin body sites evaluated. Additionally, charged spray exhibited a significantly more uniform deposition pattern as compared with similarly applied uncharged spray. Preliminary microbiological assessment documented an approximately 50-fold greater bactericidal efficacy for the electrostatic method which portends well for its further development as an effective bioterrorism countermeasure. Acknowledgements This work was supported in part by funds provided by the Georgia Agricultural Experiment Stations. Appreciation is expressed to Mr. Patrick Harrell for assistance in fabrication of experimental apparatus and graphics preparation as well as to Mr. Sean Ireland for assistance in experimental and statistical analyses. References [1] World Health Organization. 2002. http://www.who.int/health_topics/bioterrorism/en. [2] Law SE 2001 Jour. Electrostatics 51(1):25–42. [3] Law SE Cooper SC and Law WB 1999 Inst. Phys. Conf. Ser. 163:243–248. [4] Law SE 1978 Trans. ASAE. 21(6):1096–1104. [5] Law SE and Cooper SC 2001 IEEE Trans. 1A-37(6):1597–1602.

© 2004 by Taylor & Francis Group, LLC

337

Atmospheric turbulence and surface atmospheric electricity observations R G Harrison Department of Meteorology, The University of Reading, P.O Box 243, Barley Gate, Reading Berks RG6 6BB U.K. Abstract. Durable and low maintenance atmospheric electricity sensors are required if measurements which span long (year to decade) timescales are to be obtained. A passive spherical sensor, together with a high impedance electrometer, presents one method of monitoring the vertical Potential Gradient at the surface. In suitable conditions, such a sensor can monitor the state of the global atmospheric electrical system. Local factors however, such as air pollution and turbulence, influence the effectiveness of the sensor. In a series of measurements to investigate its properties, a passive sphere collector has been operated alongside (1) a long wire antenna and (2) a fine wire thermometer capable of responding to turbulent fluctuations near to the surface. In the first case the potential of the wire was found to be 3.15±0.23 greater than that of the sphere, and in the second, the average monthly Potential Gradient tended to an asymptotic value, as the magnitude of the turbulence increased.

1. Introduction Long-term monitoring of the global climate system is increasingly important, and the atmospheric electrical system is emerging as a long-neglected one aspect, which may have an effect on cloud formation [1]. Over 70years of continuous measurements of the vertical component of atmospheric electric field (or Potential Gradient, PG) at UK sites have recently been discovered [2], which provide an important resource for such studies. A long-term decline in the PG has recently been identified [3]. The UK measurements, at the Observatories of Kew (London), Eskdalemuir (Scotland) and Lerwick (Shetland), span much of the twentieth century, but ceased in the early 1980s. Although observations have continued at other sites [4], no systematic measurements in the UK have been made since then. One limitation is the availability of durable and low maintenance sensors, as it is now increasingly common for regular environmental measurements to be obtained automatically, rather than by using stations staffed with observers. A stable and well-characterised sensor is necessary if the measurements are to be used to study long-term, climate-related changes in atmospheric electricity. Measurement of the surface electric field usually use mechanical field machines [5], (field mills), or a suitable potential probe (collector) [6] connected to an ultra-high impedance electrometer [7]. Although field machines offer a more rapid time response, they eventually require maintenance because of wear to moving parts and general deterioration under atmospheric conditions. In obtaining long-term averages, reliability is generally a more important consideration than time response, and a collector-based instrument, schematically represented in figure 1, is more appropriate. The PG is derived from measurement of the potential V at the height z of the collector, and is typically ~150V.m-1 at 1m under fair

© 2004 by Taylor & Francis Group, LLC

338 weather conditions. The collector acts to acquire the same atmospheric: radioactive sources, flame probes, water droppers and long wire antennas have all been used for this purpose [5,6]. The first three such collectors all require maintenance or special precautions, and the associated apparatus may introduce geometrical distortions into the electric field. A long wire antenna produces negligible field distortion [8], but requires high-integrity insulation at each end and a careful tensioning arrangement. Only very small currents are furnished by such antenna systems, typically of order 10-14A, and electrometer instrumentation is required for the potential to be measured correctly. An alternative approach is investigated here, using a small conducting sphere as the collector rather than a long wire antenna. By using such a sphere with a diameter of a few centimetres, the collecting area is comparable with that of a long wire antenna and a similar electrometer system. This has the advantage of requiring only a single mast to support the collector, which presents some mechanical simplifications in long-term deployment and maintenance. Figure 1 illustrates the operating principle of this experimental arrangement. Near-surface turbulence mixes ions and charged particles into the region near to the sensor, which acquires the local potential of the air.

Figure 1 The atmospheric potential determined by a passive spherical sensor exposed at a fixed height connected to an electrometer voltmeter (left), and (right) practical installation of a sensor operating at z=1m.

2. Calibration In order to evaluate its operation in the atmosphere, a spherical collector was deployed alongside a long-wire antenna system [7], with both sensors at z=1m above the surface. The spherical collector was constructed from a standard table-tennis ball (radius a= 1.9cm) coated with a conductive screening paint (Nickel Screening Compound, RS part no 247– 4267). The collector was mounted at one end of a 6.25mm diameter stainless steel rod, spacing it a horizontal distance of 0.5m from a conducting vertical support mast. The spacing rod was fixed to the mast using a nylon mounting, into which the rod was fitted using a PTFE sleeve. The spherical collector assembly was positioned at the midpoint of the 20m long-wire antenna, mounted at a horizontal distance of 1m from the antenna. Both the antenna and spherical collector were connected to independent electrometers. Figure 2 shows the data obtained during a summer’s day. The wire potential Vw is always greater than the sphere’s potential Vs, but the two track each other closely. From many such measurements, it was established that the ratio k of the wire potential to sphere potential (Vw/Vs) is 3.15±0.23. when both sensors are operated at the same height. This electrostatic reduction in the © 2004 by Taylor & Francis Group, LLC

339 potential greatly relaxes the dynamic range required of the electrometer attached to the sphere, when compared with an equivalent wire, and provides an additional reason to use a sensor with simple spherical geometry.

Figure 2. Time series of potentials Vs and Vw obtained during a comparison of the passive sphere collector operating adjacent to a long wire antenna, both at 1m above the surface. The data values are 5minute averages from IHz samples.

3. Diurnal variations A failing of a wire antenna for long term use is that, depending on the sensitivity of the electrometer deployed and the conductivity of the air, physical circumstances may arise in which there is insufficient current supply to the antenna for it to function correctly. Correct operation requires turbulent mixing of natural atmospheric ions into the air adjacent to the antenna, which in turn depends on the micrometeorological conditions. In previous work [9], it has been found that turbulent mixing provides the conditions for satisfactory operation of an antenna. One method for quantifying the amount of atmospheric surface layer turbulent mixing is to use the Richardson number [10] Ri, and satisfactory operation of the passive wire has previously been found empirically to require Ri<1.0. Routine calculation of Ri requires considerably more instrumentation, however, so a rather simpler assessment of the operating regimes is desirable, preferably with only one simple additional measurement. Similar operating considerations apply to the passive sphere sensor, as the supply of current is again limited by the effective area of the sensor, and will be influenced by the charge transport from atmospheric turbulence. An investigation of these limitations was made at the same time as the calibration in figure 2 was carried out, by installing a sensitive fine wire Pt thermometer [11] close to the passive sphere sensor, figure 3. Figure 4 shows the potential on both the wire and sphere sensors as a function of time, and the temperature variations. It is clear that the potential on the wire collapses at the beginning and end of the measurements, but that the sphere shows less variation. Short-term variations in temperature do not lead to large changes in the measurements and the operation of the passive sphere cannot be reliably inferred from the rate of change of temperature.

© 2004 by Taylor & Francis Group, LLC

340

Figure 3. View of passive sphere collector operating adjacent to a fine wire thermometer, both at z=1m above the surface.

Figure 4. Variation of potential on both the long wire antenna and passive sphere sensor, during the same day in July 2002 shown in figure 2. The air temperature, measured with a fine wire thermometer, is also shown.

4. Turbulence In a longer series of measurements spanning several months at Eskdalemuir Observatory, Scotland, in the latter half of 2002, a fine wire thermometer was operated continuously alongside the passive sphere sensor. The instruments were sampled at IHz, and averages computed at 5minute intervals. Figure 5 shows the data obtained for a full day, with fair conditions for the afternoon. Comparing this data with that of figure 4, it is clear that the operation of the sensor is, again, not related to absolute temperature. Satisfactory operation of the sensor occurs when there are large and regular fluctuations occurring in the air

© 2004 by Taylor & Francis Group, LLC

341 temperature. This again indicates the importance of adequate turbulent mixing in ensuring adequate charge exchange with the sensor, as in the case of a long wire antenna.

Figure 5. Time series of air temperature and sphere potential (both at 1m), for instruments at Eskdalemuir Observatory, Scotland on 30th October 2002.

As the air temperature and its rate of change are not sufficient to identify the different operating regimes of the passive sphere, a further analysis of the effects of turbulence was undertaken. The turbulent fluctuations in temperature can also be determined from the fine wire thermometer, which has a rapid time response [11]. From micrometeorological theory, the standard deviation of air temperature σT and turbulent temperature scale T* are related by (1) where ψ is a semi-empirical function of the sensor height z, the local displacement height d and the Obukhov Length L, a function of atmospheric stability [10]. σ T therefore provides an indication of the turbulence. Turbulence considerably varies between strong mechanical and thermally generated turbulence in daytime (unstable) conditions, to weak turbulence in nocturnal (stable) conditions. Standard deviations of temperature σ T were calculated for 15minute periods, again using 1Hz data. The average sphere potential measured in the same interval was plotted against the temperature standard deviation, for a set of measurements extending over one month, Figure 6. It is clear that sphere’s potential is sensitive to the amount of turbulent mixing and that σ T does provide a readily measured indicator of the controlling turbulence scales. Using a logarithmic fit, the asymptotic value of the equivalent potential at 1m can be estimated.

© 2004 by Taylor & Francis Group, LLC

342

Figure 6. The average equivalent potential at 1m (kVs) using data from one month, found by selecting the values of Vs when the surface layer turbulence, as measured by the standard deviation of temperature σT, is greater than the specified minimum value on the x-axis.

5. Discussion Long term measurements of atmospheric potential gradient, derived from a point measurement of potential require simple and low maintenance instrumentation. A passive sphere sensor, connected to a sensitive electrometer, provides measurements of potential when there is sufficient turbulent mixing for sufficient charge exchange with its surroundings to occur. As a result it will only operate for a fraction of the day, however this will not present a significant problem if a long-term average is sought. The practical requirement of fair weather conditions is another limitation always present, and therefore monthly averages are typically derived. Measurements of standard deviations of temperature have been demonstrated to provide an independent assessment of the operation of the sensor, although further characterisation of the micrometeorological effects is desirable. Acknowledgement Assistance was given by the British Geological Survey and Met Office at Eskdalemuir. 6. References Carslaw K S, Harrison R G and Kirkby J 2002 Science 298, 5599, 1732–1737 Harrison R G 2003 Weather 58, 11–19 Harrison R G, 2002 Geophys Res Lett, 29(7) DOI 10.1029/2002GL014878 Märcz F 1990 Ann. Geophysicae 8, 525–530 Chalmers J A 1967 Atmospheric Electricity, (2nd edition, Pergamon Press, New York,) Israel H 1973 Atmospheric Electricity (Israel Programme for Scientific Translations) Harrison R G 1997 Rev. Sci. Inst 68, 3, 1599 Crozier W D 1965 J. Geophys. Res. 68, 3451 Barlow J F and Harrison R G 1999 In Christian H J (ed) Proc 11th International Conference on Atmospheric Electricity, Guntersville, Alabama (NASA/CP-1999–209261, 575,) [10] Kaimal J C and Finnigan J J 1994 Atmospheric boundary layer flows: their structure and measurement (Oxford University Press) [11] Harrison R G and Pedder M A 2001 Rev Sci Inst, 72, 2, 539

[1] [2] [3] [4] [5] [6] [7] [8] [9]

© 2004 by Taylor & Francis Group, LLC

343

Dust particles removal by a novel two-stage electrostatic precipitator A Jaworek1, A Krupa1, K Adamiak2 1 Institute of Fluid Flow Machinery, Polish Academy of Sciences, Fiszera 14, 80– 952 Gdañsk, POLAND 2 Department of Electrical and Computer Engineering, University of Western Ontario, London, Ontario, CANADA N6A 5B9 Abstract. A two-stage electrostatic precipitator is a gas-cleaning device in which the charging and precipitation processes are separated. This device can be used to avoid back-corona discharge, when dust of high resistivity has to be removed from an exhaust gas. A bench-top size laboratory model of the two-stage electrostatic precipitator was investigated experimentally. The charging stage was accomplished using the alternating electric field charger, a device that allows charging the particles to higher levels than conventional DC-corona chargers. The precipitation stage was created by a set of five parallel plates, two of which were maintained at high potential and the remainder grounded.

1. Introduction In order to avoid difficulties in collecting fly ash of high electrical resistivity, the charging and collection processes can be separated and accomplished in two different stages. Such a device is known as two-stage electrostatic precipitator. First, the particles are passed through the charging stage, where they are charged in a mono-ionised electric field, similar to a conventional electrostatic precipitator, where the voltage has to be reduced to eliminate the back-corona discharge. Next, the particles enter the precipitation stage, which is usually formed by a set of parallel plates with every second plate maintained at high potential and remainder grounded. Such a set of electrodes is free of the back-corona discharge due to lack of the ionic current. The two-stage electrostatic precipitator was considered first by Masuda and Hosokawa [1] to control the emission of high resistivity dust. The boxer charger was used in their experiments as the charging stage [2,3]. The collection section was made of 4 electrodes spaced at 150 mm. Different types of precharger were developed and tested in two-stage electrostatic precipitators since: a nozzle charger equipped with corona electrodes, through which the particles flowed with high velocity to prevent them from precipitation [4], coronatriode charger [4], and a quadrupole precharger with four passive rods placed at a square corners and a discharge electrode at the centre of this arrangement [5]. In this paper, the alternating electric field charger [6–8] is proposed and tested for the first time as a precharger in a two-stage electrostatic precipitator. The particles are charged by an ionic current in the alternating electric field, and the charge imparted to the particles is close to the Pauthenier limit, i.e., is much higher than that available in other techniques. Regardless of numerous investigations on the two-stage electrostatic precipitators, many questions remained unanswered. One issue is the re-entrainment of the dust particles from

© 2004 by Taylor & Francis Group, LLC

344 the collection electrodes due to lack of cohesive forces, which are present in standard electrostatic precipitator because of the ionic current flowing through the dust layer. This problem was, hitherto, not investigated and is considered in this paper. 2. Experimental The schematic view of experimental model of the two-stage electrostatic precipitator with alternating electric field precharger is shown in Fig.1. The precharger and collection electrodes were mounted in a duct of square cross section of 160×160 mm, made of clear acrylic plastic. The precharger consisted of two parallel grids made of 9 stainless steel rods, 2 mm in diameter spaced 20 mm from each other, and two discharge electrodes placed outside of the grids. Each discharge electrode was made of a brass sheet with 98 stainless steel pins 0.6 mm in diameter and 5 mm long, protruding 5 mm from the sheet. The distance between the grids was 50 mm, and from each grid to the tip of the pins behind 45 mm. The precharger electrode system was connected to two high voltage transformers (ZWAR-ABB, Poland, model UMZ-24) connected in anti-phase, through the circuit made of diodes and resistors, as shown in Fig.1. The ionic charging current is emitted by the discharge electrodes, and flows through the grids to the charging zone between the grids. The alternating electric field existing within the charging zone causes the particles to oscillate, but prevents them from migrating to the charger electrodes, thus decreasing their precipitation at this stage.

Fig.1. Schematic diagram of the experimental set-up of two-stage electrostatic precipitator with the alternating electric field precharger (not in scale).

The electrodes in the collection section were made of a brass sheet with 1 mm thickness. Five parallel plates 130 mm high and 160 mm long downstream were placed in the duct. The space between them was 30 mm. Two plates were maintained at high potential from the high voltage power supply (SPELMANN HV model SL600W/30kV/P), and other three were grounded. A flow straightener was mounted at the inlet of the channel to eliminate the air vortices. At the outlet of the duct a collection grid, made of fine phosphor-bronze mesh of holes of about 40 µm, was used to measure the electric charge carried by dust particles, which escaped the precipitator. Dust from the last stage of an electrostatic precipitator installed in the Lambton Power Station, Ontario, Canada of mean particle diameter of about 10 µm was used in the experiments. The particles were fluidised by compressed air and injected into the channel, where they were charged by the alternating-electric-field charger. The airflow through the

© 2004 by Taylor & Francis Group, LLC

345 duct was forced by a suction fan placed at the outlet of the duct. The trajectories of the particles in the collection stage were recorded with a CCD digital camera (SONY DCRTRV525). The space between the collection electrodes of the precipitator was illuminated by a light beam of about 2 mm in thickness. The gas velocity in the channel was measured by a hot wire anemometer (Mini Anemometer Series 490, Kurtz Instruments Inc.). The size distribution of the particles was measured by Malvern Particle Size Analyser model 2600. The Sauter mean diameter of the particles D3,2 was equal to about 10 µm (Fig.2). The dust volume concentration within the duct measured by the Malvern was in the range from 0.0004 to Fig.2. Fly ash particle size distribution 0.0014 %. 3. Experimental results The obscuration was measured behind the collection section by means of the Malvern Particle Size Analyser. The obscuration ratio, shown in Fig.3, is the relation of obscuration measured for the voltage applied to the collection section, to the obscuration when the voltage was off and the dust could freely flow through the precipitator. Because the fluidiser does not work smoothly and the particle concentration changes with time, the switch-on/ switch-off procedure was repeated a few times, and the ratio in Fig.3 is a mean value of the obscurations measured during these periods. The obscuration ratio decreases with increasing voltage applied to the collection electrodes and ratio can be a measure of the precipitator collection efficiency. The collection efficiency can be estimated to 60–70% for fly ash particles, and over 90% for MgO, if the voltage is sufficiently high.

Fig.3. Obscuration ratio at the outlet of the two-stage electrostatic precipitator for fly ash (a) and MgO (b) particles.

The particle migration velocity to the collection electrodes results from the balance of Stokes drag force: Fs=πη6rvy (1) and electric field force: Fe=QE (2) where: Q is the particle charge, E the electric field between the electrodes, η-the gas viscosity, and r -the particle radius.

© 2004 by Taylor & Francis Group, LLC

346 Assuming the AC charger is supplied by a voltage of 14 kVrms, the Pauthenier charge limit for the fly ash particles of 10 µm in diameter is 2.3 fC (assuming that its dielectric constant is that of SiO2—the main component of the fly ash), and for MgO is 2.73 fC. The velocity, transversal to the gas flow, for the particle charged to the Pauthenier limit is 0.38 m/s and 0.54 m/s, respectively, for the collection electrode voltage equal to 12 kV. This means all particles should be collected over the first few centimetres of the collection section. The problem of particle re-entrainment was also studied. It was noticed that at higher voltages (10–12 kV) the fly ash particles begin to jump between the collection electrodes. Examples of the dust particle motion between the electrodes are shown in Figs.4 and 5. Each figure is a combination of a few consecutive frames of the video recordings. The particles of MgO (Fig.5) detach as agglomerates, while the fly ash particles are much numerous, but because they are single they are hardly visible (Fig.4). The electric force is sufficiently high to overcome the cohesion forces and some particles are removed from the dust layer due to charge obtained by conduction. These particles flow to the opposite electrode, where they can be recharged. After a few ‘jumps’ they can ultimately leave the interelectrode space. The re-entrainment of the particles is also evidenced by increasing obscuration measured behind the collection stage. Using the trajectories of the particles reentrained from the collection electrodes their charge can be estimated using the equation: (3) where α is the angle between the particle trajectory and the gas velocity vector, d—the electrode spacing. It is assumed that no cluster is detached from the collection electrodes, but only a single fly ash particle of mean size 10 µm. For these particles the charge Qr is 1.4 fC. The phenomena are quite different for particles of high cohesiveness, like MgO. The particles are easily agglomerated and cannot be re-entrained even for higher voltages, close to the discharge onset level. Only rare jumps were observed, and usually large agglomerates were involved (Fig.5). The particle re-entrainment did not have a noticeable effect on the current, probably because the number of both positively and negatively charged re-entrained particles was balanced.

Fig.4. Trajectories of re-entrained fly ash particles between collection electrodes.

Fig.5. Trajectories of MgO particles between the electrodes of the collection section.

From the recorded trajectories of the particles between the collection electrodes it can be seen that the dust particles, or dust cluster, detach from the electrode due to the electrostatic force as soon as it is re-charged to the opposite polarity. Then, they flow to the opposite electrode. After impact on this electrode the following situations are possible:

© 2004 by Taylor & Francis Group, LLC

347 1. 2. 3.

The particle rebounds off the electrode, switching the polarity of its charge, and with opposite charge returns to the flow. The particle can also be recharged by oppositely charged particles and as neutralised returns to the flow, as shown in Fig.5; The particle rebounds off the electrode with opposite charge, and returns directly to the opposite electrode; The particle impacts the electrode and sputters a new particle oppositely charged, but remains itself on the electrode surface.

Fig.6. Current ratio at the precipitator outlet vs. collection electrode voltage for fly ash particles

The current at the precipitator outlet was also determined to estimate charged particle concentration leaving the precipitator. The ratio of the currents flowing to the collection grid with and without voltage applied to the collection electrodes for the fly ash particles is presented in Fig.6. The current ratio decreases with increasing voltage, but for the voltage above about 14 kV it sharply increases because of an electrical discharge from the collection electrodes.

4. Discussion Two-stage electrostatic precipitators were proposed 20 years ago to increase the collection efficiency of fine particles, especially with high resistivity, and to prevent back-corona discharge. However, many experiments show (cf. Table 1) that this is not a highly effective device. Usually, the collection efficiency is lower than 90% mainly because of, as it was shown in our experiment, low cohesiveness of the collected dust. Higher collection efficiencies were obtained only for highly cohesive particles, mainly DOP (cf. Table 1) or with simultaneous spraying of the water aerosol (Yoo et al. [13]).

Table 1. Two-stage electrostatic precipitators—comparison of the results

© 2004 by Taylor & Francis Group, LLC

348 From the previous experiments it can be also deduced, and this remains in agreement with our results, that the method of particle charging is less important than the particle collecting technique. Bart et al. [10] subjected the particles to oscillatory motion in order to promote their agglomeration in the collection section by using AC electric field, but the collection efficiency in this type of precipitator was not very high. Chang et al. [5] succeeded with high collection efficiency for particles of diameter larger than 0.1 µm, because the collection stage was of the wire-plate geometry with discharge current reduced to prevent the backcorona discharge. However, no significant effect of precharging on the collection efficiency was observed for particles smaller than 0.1 µm. 5. Conclusions From the experimental results presented in this paper it can be concluded that the cohesive forces play a fundamental role in two-stage electrostatic precipitators. For dust of high cohesiveness (MgO) the collection efficiency is very high and only seldom re-entraining particles were observed. For particles of low cohesiveness (fly ash) strong re-entrainment was recorded, which was mainly caused by the electric forces in the collection stage. A charge conducted to a particle from the electrode causes this particle to be detached from one electrode and attracted to the opposite. The zigzag motion of particles between the collection electrodes and their flow off the precipitator were observed. This implies that the charge on the particle is the same but of opposite polarity and is acquired by induction processes. However, the mechanism of dust particle rebounding is unclear. The angles of impact and rebounding are usually equal and full understanding of the role of mechanical and electrical forces requires further investigations. Two-stage electrostatic precipitation can be effective only if the cohesiveness of the dust will be increased. Moisturising of the dust layer may be one of possible remedies. Acknowledgements. This paper has been supported by the Natural Sciences and Engineering Research Council (NSERC) of Canada. The authors wish to thank Mr. Randall M. Van Hooren from the Ontario Power Generation for providing the precipitator dust samples. The authors are grateful to Mr. Marian Jaworski from the Department of Mechanical and Materials Engineering, University of Western Ontario for his technical help, which have made these investigations possible. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

Masuda S., Hosokawa Sh., IEEE Ind. Appl. Soc. Cof. Rec. (1982), 1094–1101 Masuda S., 2nd Int. Conf. Electrostatic Precipitators, Kyoto, Nov. 1984, 177–85 Masuda S., Washizu M., Mizuno A., Akutsu K., Conf. Electrostatic Precipit., Leura, 21–24 Aug. 1978 McLean K.J., IEE Rev. 135 (1988), Pt.A, No.6, 347–361 Chang J.S., Looy P.C., Webster C., Berezin A.A., Zukeran A., Ito T., 7th Int. Conf. Electrostatic Precipitation. 20–25 Sept 1998, Kyongju, 620–627 Jaworek A., Krupa A., J. Electrostat. 23 (1989), 361–370 Adamiak K., Krupa A., Jaworek A., Inst. Phys. Conf. Ser. No. 143, Bristol 1995, 275–278 Lackowski M., J. Electrostat. 51–52 (2001), 225–231 Hautanen J., Janka K., Keskinen J., Lehtimäki M., Kivisto T., J. Aerosol Sci. 17 (1986), No.3, 622–626 Bart S.F., Melcher J.R., Ehrlich R.M., Ind. Eng. Chem Res. 27 (1988), No.l, 123–131 Asano K., Higashiyama Y., Yatsuzuka K., Nishimura R., 6th Int. Conf. Electrostatic Precipitation. 18–21 June 1996, Budapest, 57 Asano K., Choi Ch., Yatsuzuka K., Lim H., Polish-Japanese Symp. Non-Thermal Plasma Processing of Water and Air. 29–31 May 2000, Sopot, 101–104 Yoo K.H., Lee J.S., Oh M.D., Aerosol Sci. Technol. 27 (1997), No.9, 308–323 Yoo K.H., Lee H., Choi M., Lee J.S., Oh M.D., 7th Int. Conf. Electrostatic Precipitation. 20–25 Sept 1998, Kyongju, 93–98 Kawada Y., Jindai W., Zukeran A., Ehara Y., Ito T., Takahashi T., Takamastu T., 7th Int. Conf. Electrostatic Precipitation. 20–25 Sept 1998, Kyongju, 76–83 Kubo T., Kawada Y., Zukeran A., Ehara Y., Ito T., Takahashi T., Kawakami Takamatsu T., J. Aerosol Sci. 30 (1999), Suppl.l, 793–794

© 2004 by Taylor & Francis Group, LLC

349

Dust particle removal by wet-type electrostatic scrubber A Krupa1, A Jaworek1, T Czech1, M Lackowski1, J Luckner2 Institute of Fluid Flow Machinery, Polish Academy of Science P.O. Box 621, PL-80952 Gdansk, Poland, [email protected] 2 Military Institute of Chemistry and Radiometry Al. Gen. A. Chrusciela 105, 00–910 Warsaw, Poland 1

Abstract. Cleaning of exhaust gases from micrometer size particles by charged aerosol was investigated experimentally. The aerosol was generated by electrospraying using a multinozzle spray system. The dust particles were charged in a dc corona discharge charger. The dust was deposited on the droplets due to Coulomb forces. The collection efficiency for fly ash particles of mean size of 5 µm was higher than 90%, and water consumption lower than 0.08 1/m3.

1. Introduction The efficiency of removal of dust particles smaller than a few micrometers in diameter by conventional electrostatic precipitators decreases with decreasing particle size. The electric charge on such particles cannot be too large and, as a result, the air drag force can be stronger than the electric force that causes the particles to pass the precipitator. Additionally, small fly ash particles can be easily re-entrained from the collection electrode because the cohesive forces are very weak. Conventional wet scrubbers are also ineffective in removal fine particles from exhaust gases due to low inertial forces on the dust particles, which can flow around the droplet following the stream lines. The removal of small particles can be enhanced when the dust particles and the scrubbing droplets are electrically charged to the same or opposite polarities [1–5]. The charged droplets sweeping the precipitation chamber act as small collectors, attracting the particles due to Coulomb force. This method can be of much higher overall collection efficiency as compared to the conventional electrostatic precipitators or inertial scrubbers. This paper presents the experimental results of measurements of the collection efficiency of a laboratory scale electrostatic scrubber. Usually, mechanical (rotary, pneumatic) atomisers with induction charging were tested as a source of charged droplets. In this paper we use an electrohydrodynamic atomiser for generation charged spray. 2. Experimental The measurements were carried out in the experimental system shown schematically in Fig.1. The channel of rectangular cross section of 160×160 mm was made of plexi glass. The particles were charged in a corona discharge between a needle electrode and a grid. The discharge electrode was made of a stainless steel needle 0.9 mm in diameter and 5 mm long. The grid was made as a set of stainless steel rods 2 mm in diameter, and spaced at 20 mm

© 2004 by Taylor & Francis Group, LLC

350 each other. The distance between the needle tip and the grid was 20 mm. The needle was at the potential of -12 kV or +13 kV relative to the grid. The ionic current emitted by the discharge electrode flows to the grid. The dust particles, which are injected through 6 holes placed around the needle, are charged by this ionic current.

Fig. 1. Schematic view of an experimental arrangement for electrostatic scrubbing investigations.

An electrospray system was used as a source of charged droplets. The spray system was constructed from four stainless steel capillaries 0.5 mm i.d., 0.7 mm o.d, and 12 mm long, and grounded electrodes placed in the plane of the nozzle tips (Fig.2a). The ground electrodes were ring-shaped, and they were made of brass wire 3 mm in diameter. The inner diameter of each ring was 28 mm. The distance between the spray nozzles was 30 mm. The nozzles were connected to a high voltage power supply (SPELMANN HV model SL600W/30kV/P). A photograph of the spray plume generated by the system, taken out of the channel, is presented in Fig.2b. The droplets size distribution was measured by an optical particle size analyser, model AWK made by K&K (Poland). The voltage at the capillaries was adjust to the magnitude at which the precession mode (cf.[6]) of spraying was obtained. This mode is characteristic in that the whole jet rotates regularly around the capillary axis, and its end also rotates around the mother jet. For this mode, the spray plume is most wider and the droplets are relatively small, that can be advantageous for the scrubbing applications. The size distribution of distilled water droplets is shown in Fig.3. The number and volume mean sizes were of about 80–90 µm in diameter (Fig.3). At the outlet of the duct a collection two-layer paper-fibrous filter was used to collect the dust escaping the scrubber. The filter was equipped with a metal grid made of fine mesh to measure the charge carried by dust particles. The mesh was made of phosphor-bronze wire with holes of about 40 µm. At the inlet of the channel a flow straightener was mounted to eliminate the vortices. The air flow velocity was controlled by the hot wire anemometer TSI model 8455. A grounded guard electrode was placed between the spray system and the collection grid to eliminate the leakage current from the high voltage electrodes, which could disturb the measurements of the current from the grid.

© 2004 by Taylor & Francis Group, LLC

351

Fig.2. Scheme of the four-nozzle spray system (a), a photograph of the spray plume of the fournozzle spray system (voltage to the nozzles 10 kV) (b).

Fig.3. Number and volume size distributions of the spray generated in the four-nozzle spray system for voltage of 10 kV to the nozzles.

Dust from the last stage of an electrostatic precipitator of mean particle size of about 5 µm was used in the experiments. The dust size distribution was determined using Laser Aerosol Particle Size Spectrometer model LAP 320 of TOPAS GmbH. The mean size of dust collected on the filter at the duct outlet, i.e., that escaping the electrostatic scrubber was 6 µm. The size distribution of both samples of the dust are shown in Fig.4. The dust/air mixture was generated in a specially designed fluidised bed by compressed air, and injected into the channel through holes in the corona charger. The air flow through the duct was forced by a suction blower placed at the outlet of the duct.

© 2004 by Taylor & Francis Group, LLC

352

Fig.4. The size distribution of the dust used in the experiments and that escaping the scrubber.

3. Experimental results The collection efficiency of the electrostatic scrubber was estimated by weighting the dust injected to the scrubber and that collected on the paper tissue. The total mass min of the dust injected to the scrubber is the difference between the dust weight in the fluidiser before and after each run of the experiment. The mass mcoll of the dust collected by the paper tissue was determined by weighting the tissue before and after the run. Two consecutive runs, 6 minutes each, were necessary for estimation the collection efficiency. The first run was for the scrubber operating without the spray and corona charger, and the second one with the charger and spray system turned on. The collection efficiency was estimated from the equation:

(1) where: mcoll/on and mcoll/off is the mass of the dust collected on the tissue with and without the spray, respectively, and min/on and min/off is the mass of the dust injected to the channel, for each case respectively. The dust loading in the experiments was about 1 g/m3, and mean particle concentration was estimated to the order of magnitude of 109 particles per cubic meter. The water consumption ranged from 0.03 to 0.08 1/m3 depending on gas velocity. The specific charge of the dust particles estimated from the total current from the metal mesh and the mass of the collected dust varied between 10 and 20 mC/kg, irrespective of gas velocity. The variations in specific charge were probably caused by changes in spraying and charging conditions, including changes in air humidity and uncontrollable corona discharge current or tribocharging. The charge on a particle of mean size of 6 µm determined from the measurements of the current collected by the mesh at the channel outlet can be estimated to 2–4 fC. The collection efficiency determined in these experiments is shown in Fig.5. It can be noticed that charging the droplets and dust particles result in more efficient removal of the particles by the scrubber. Charging the droplets and the dust to opposite polarities allows the particles to be collected on the droplets while traversing the aerosol shower. Charging the dust and droplets to the same polarity cleans the gas more effectively than opposite charging. This result can be explained by repulsion of dust particles by the aerosol shower,

© 2004 by Taylor & Francis Group, LLC

353 and precipitation them on the duct walls. The cleaning of the gas from fine particles is more efficient for lower velocity. The collection efficiency by charged aerosol for uncharged dust is also shown in Fig.5 for reference. The collection efficiency drops very quickly with gas velocity increasing. Further increase of the collection efficiency can be achieved by using a two-stage scrubber, that however, results in higher water consumption.

Fig.5. Collection efficiency of the electrostatic scrubber.

4. Conclusions The results of experimental investigation of removal of dust particles by electrostatic scrubber have been presented in the paper. The results show that removal of dust particles by water droplets can be increased by electrical charging of the droplets and particles to the same or opposite polarities. The equipment utilising electrostatic forces operates more effectively for lower gas velocities than that in which inertial collection is dominant. The water consumption in this type of cleaning device is also very low, it can be lower than 100 ml/m3, that is much lower than that in conventional inertial scrubbers. The collection efficiency was up to 90% for dust and droplets charged to the same polarity and gas velocity of about 0.2 m/s. This method can be applied for industrial or flue gas cleaning. In an industrial scale equipment the spray system could be installed in wider channel but of similar height. Acknowledgements This paper has been supported by the State Committee for Scientific Research of Poland (KBN Grant No. 1440/T10/2000718). References 1.

Kraemer H.F., Johnstone H.F., Collection of Aerosol Particles in Presence of Electrostatic Fields. Ind. Eng. Chem. 47 (1955) No 12, 2426–2434

© 2004 by Taylor & Francis Group, LLC

354 2. 3. 4. 5. 6.

Schmidt M., Theoretische und experimentelle Untersuchungen zum Einfluss elektrostatischer Effecte auf die Nassentstaubung. Dissertation. Universitat Karlsruhe 1993 (in German) Metzler P., Weiss P., Büttner H., Ebert F., Electrostatic Enhancement of Dust Separation in a Nozzle Scrubber. J. Electrostatics 42 (1997), No. 1–2, 123–141 Jaworek A., Krupa A., Adamiak K., Submicron Charged Dust Particle Interception by Charged Drops. IEEE Trans. Ind. Appl. 34 (1998), No.5, 985–991 Adamiak K., Jaworek A., Krupa A., Deposition Efficiency of Dust Particles on Water Droplets in Electrostatic Scrubbers. IEEE Ind. Appl. Soc. Annual Meeting, St. Louis, Oct. 12–15, 1998, 1919–1926 Jaworek A., Krupa A., Classification of the modes of ehd spraying. J. Aerosol Sci. 30 (1999), No.7, 873–893

© 2004 by Taylor & Francis Group, LLC

355

Expert System Applications for Electrostatic Separation Processes Lucian Dascalescu,1 Adrian Samuila,1 Michaela Mihailescu,2 Alexandru Iuga,2 Reiner Köhnlechner3 1 Electronic&Electrostatics Research Unit, LAII-ESIP, EA 1219 University Institute of Technology, 4, av. Varsovie, 16021 Angouleme Cedex, France 2 Technical University, 15, C. Daicoviciu street, 3400 Cluj-Napoca, Romania 3 Hamos GmbH, 17 im Thai, Penzberg, Germany Abstract—This paper aims to analyze the structure of an expert system that has been developed in order to capture expertise in the field of electrostatic separation processes and to make it available to all those involved in their industrial application. A knowledge base was created and is continuously updated. In addition to the basic rules that were formulated for guiding the choice of the most appropriate electrostatic separation technique, an experimental design assistant has been integrated in the expert system. This enables tuning of the operating parameters in accordance with the characteristics of the granular mixture to be processed. The second block of the expert system proved to be effective in transferring the expertise in trouble shooting and diagnosis from research or supervisory staff to process operators.

1. Introduction The application of expert systems for process diagnosis, operator support and training in mineral processing, as well as in chemical and metallurgical industry, started in the early 1980’s [1]-[6]. By the end of the 1990’s, electrostatic separation techniques had become a mature field of engineering [7–11], which justified the authors’ attempt to prototype a dedicated expert system [12]. The objective was to capture expertise and to make it available to all those involved in the industry application of these techniques. The development of the expert system applications took into account the various methods of knowledge acquisition described in the literature [13–15]. The aim of the present paper is to analyze the peculiarities of that expert system for electrostatic separation processes. Three applications are discussed: (1) choice of equipment; (2) optimization of the operating parameters; (3) trouble-shooting guide. 2. Representation of knowledge on electrostatic separation processes The key issue of knowledge representation consists in the definition of the structure employed to organize the acquired expertise in a form suitable for computer processing [16]. The choice of MERISE formalism [17] for representing the knowledge on electrostatic separation technology and equipment was related to the peculiarities of the two applications that were addressed at this stage of development of the expert system.

© 2004 by Taylor & Francis Group, LLC

356

Fig.1. MERISE conceptual model of the knowledge on the composition of the various granular mixtures, on the types of electrostatic separators and the respective electrode configurations.

The first application that was analyzed focused on the choice of the most appropriate electrostatic separation equipment for the processing of a given granular mixture. The corresponding MERISE conceptual model reflects the relations that exist between six types of entities: ‘MIXTURE’, ‘CONSTITUENT’, ‘MATERIAL’, ‘SIZE CLASS’, ‘SEPARATOR’ and ‘ELECTRODE’. Each entity has several properties, one of them serving as identifier. The examples in Table 1 show that each occurrence of the entity ‘CONSTITUENT’, for instance, is characterized by a different value of the identifier ‘CodeConstituent’. Each occurrence of the entity ‘MATERIAL’ of which a ‘CONSTITUENT’ is «MADE OF » has a different identifier ‘CodeM’ and contains a list of physical properties that are relevant for the electrostatic separation techniques. These values are either those found in the literature or measured by the researcher (for new types of plastics or new minerals). One of the difficulties in the development of this part of the expert system is the multitude of configurations that should be taken into account. The example given in Fig. 2 illustrate the way in which the significance of the properties ‘Diml’, ‘Dim2’, … of the entity ‘ELECTRODE’ have been defined in connection with the ‘Type’ of the electrode that was considered for that application: DUAL WIRE, … Similarly, the properties ‘Posl’, ‘Pos2’, … of «CONFIG» define the position of the electrode by taking into account the ‘Type’ of the ‘SEPARATOR’ (which is CORONA with ROLL CARRIER ELECTRODE) and the ‘Type’ of the ‘ELECTRODE’: ‘Pos1’=␣, ‘Pos2’=h, ‘Pos3’=␤.

Table 1 Examples of occurences of the entity ‘CONSTITUENT’

© 2004 by Taylor & Francis Group, LLC

357

Fig. 2. Schematic representation of the electrode system of a roll-type electrostatic separator.

The relation PROCESSED associates a given ‘MIXTURE’ to an existing ‘SEPARATOR’, while the relation «CONFIG.» defines the corresponding ‘ELECTRODE’ arrangement. For instance, the ‘MIXTURE’ encoded RECYCL01 can be PROCESSED on a ROLL ‘Type’ ‘SEPARATOR’, ‘CONFIG.’-urated as in Fig. 2. In practice, the user has to indicate to the expert system the properties of the ‘CONSTITUENT’ of a ‘MIXTURE’ to be PROCESSED and the associated ‘SizeClass’. If these information correspond to a ‘MIXTURE’ that is already in the knowledge base, the type of recommended ‘SEPARATOR’ and of its optimal ‘ELECTRODE’ configuration are indicated. In all the cases when the characteristics of the solids to be separated in a new application do not correspond to any of the repertory mixture, the expert system is able to guide the user through the selection process, following the rules synthesized in Table 2. Thus, for the separation of a granular mixture containing large insulating particles, the expert system recommends the use of two corona electrodes.

Table 2 Selection of the electrostatic separator for a given granular mixture

© 2004 by Taylor & Francis Group, LLC

358

Fig. 3. Recommended operating parameters for the electrostatic separation of a 50% Cu—50% PVC granular mixture (chopped electric cable and wire wastes).

Regarding the second application targeted by the present work, namely the prescription of optimum operating parameters, the first prototype of the expert system [12] enabled only the fitting between a new application and a similar electrostatic separation process already existing in the knowledge base (Fig. 3). Taking into account the wide variety of different materials that have to be treated by an electrostatic separator operating in an industrial environment, this part of the system proved to be ineffective if not associated with a program capable to guide the optimization of the process by using experimental design techniques. In its present form, the expert system provides recommendations on the domain of variation of the various factors and constraints to an independent program of experimental design [18]. This program proposes the electrostatic separation tests that should be carried out for identifying the model of the process and the optimum setting of the input variables (high voltage level, position of the electrodes, feed rate, etc.). Hamos GmbH, Penzberg, Germany, is currently using this procedure for establishing the optimum operating parameters for the newly developed electrostatic separation technologies. 3. Trouble-shooting and diagnosis of electrostatic separation processes For the expert system to be able to diagnose the operation of an electrostatic separation process, it has been provided with a list of symptoms as the starting point for the knowledge base query aimed at the identification of the problem that might have generated the malfunction. The elaboration of rule sets on which the most probable diagnostic may be established for each symptom has been an extremely complex activity. The data regarding the failure histories of several types of industrial electrostatic separators have been collected and analyzed. The conclusions have been corroborated with the knowledge on the physical phenomena that govern the operation of the respective type of separator. The diagnostic and trouble-shooting procedures have finally been validated by a 6-month “in-field” survey of an industrial process. The example given in Fig. 4 illustrates this approach in the case of industrial roll-type corona-electrostatic separators employed for the processing of chopped electric wire and

© 2004 by Taylor & Francis Group, LLC

359 cable wastes. More than 75% of the malfunctions registered for this type of application are evidenced by the fact that the metal collected in the dedicated compartment has lower purity than the one expected under normal operating conditions. This symptom indicates a relative decrease of the “pinning effect”, due to the image force Fi, in comparison with the “detaching effect”, due to the centrifugal force Fc. As Fi is proportional with the square of the electric charge Q, the value of which depends on the strength of the corona field, the first action to take is check if the supply voltage U of corona electrodes has the prescribed value. The average value of the voltage Uav across the inter-electrode gap may decrease if the frequency of spark discharges exceeds the upper recommended limit. These discharges are normally caused by conducting particles passing through the electric field and which reduce the dielectric strength of the air gap. This symptom can indicate either that the concentration of metal particles in the granular mixture is higher than presumed, or that their dimensions are larger than the prescribed values. A detailed explanation of this and other symptoms is presented elsewhere [12]. The survey of an industrial process at RIPS-ALCATEL, France, indicated that in 42% of the cases the malfunction was due an accidental modification of the size characteristics of the constituents (particle sizes are larger than the ones taken into account when the optimum parameter values were established), in 36% of the cases the disturbing factor was a change in the conductivity of insulating particles (material humidity is higher than the normal value; conditioning temperature of the granular mixture is lower than the prescribed value), in 15% of the cases an increase of the material feed flow was incriminated (the mono-layer presentation of the material on the surface of the cylindrical electrode is no more assured), all the other causes representing only 7% of the registered events. In each of the above-mentioned cases, the operator followed the advice of the expert system: correct the operating parameters of the grinders; increase the temperature at which the drying or the thermal conditioning of the material is accomplished; diminish the material feed flow. For the considered process, the expert system was able to indicate the values at which the respective parameters should be set. In other situations, it is very likely that statistical process control and experimental tools should be employed for effective diagnosis and trouble-shooting. Efforts are presently made to integrate this kind of tools in this block of the expert system.

Fig. 4. Expert system diagnosis of a metal purity problem.

© 2004 by Taylor & Francis Group, LLC

360 4. Conclusions The experience accumulated by the authors since 1999, when the present work was started, suggests that the expert systems are likely to become the next wave of computer applications to hit the mature electrostatic technologies, such as precipitation of dust, spraying of powders, or sorting of granular mixtures. The industry application of the prototype of such an expert system for electrostatic separation processes clearly has demonstrated how efficiently the knowledge of many experts can be integrated and efficiently stored, in a flexible structured data base. The modularized nature of the rule-based expert system enables the continuous updating of the knowledge base. The learning tools of the expert system contribute to a diminution of the training costs of the new operators for industrial electrostatic separators. The efficient trouble shooting assistance provided by the expert system offers faster solutions for most of the current operation problems. Even if the advantages of such an expert system have been clearly demonstrated, and most of the difficulties related to the development of a fully-operational expert system for electrostatic separation techniques have been surpassed, further research is needed to expand the field of expertise covered by this tool. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

Weiss S M and Kulikowski C A 1984 A Practical Guide to Designing Expert Systems (London: Chapman and Hall). Hayes-Roth F (Ed) 1986 A Guide to Expert Systems (Ontario: Addison-Wesley). Harmon P, Maus R and Morrissey W 1988 Expert Systems. Tools and Applications (New York: Wiley). Benchimol G, Levine P and Pomerol J C 1990 Systemes experts dans I’entreprise. (Paris: Edition Hermes) Laguitton D and Leung J 1989 Advances in expert system applications in mineral processing in Processing of Complex Ores (New York: Pergamon) 565–574 Proceedings of AIPAC’89 (Advanced Information Processing in Automatic Control) 1989 (Oxford: Pergamon) Inculet I 1986 Electrostatic Mineral Separation (New York: Wiley) Higashiyama Y and Asano K 1998 Paniculate Science and Technology 16 77–90 Knoll F S and Taylor J B 1985 Mineral and Metallurgical Processing 2 106–114 Dascalescu L, Morar R, Iuga A, Samuila A and Neamtu V 1998 Paniculate Science and Technology 16 25–42. Kwetkus B A 1998 Paniculate Science and Technology 16 55–68 Mihailescu M, Samuila A, Iuga A and Dascalescu L 1999 Conf. Rec. IEEE/IAS Ann. Meet. (Pheonix: IEEE Press) Frenzel L E 1987 How to develop an expert system in MacLaren D (Ed) Understanding Expert Systems (Indianapolis: Howard W. Sams & Co) 149–182 Hart A 1986 Knowledge Acquisition for Expert Systems (Toronto: McGraw-Hill). Harris C A, Woo E, Hall M and Miller R 1989 Knowledge acquisition techniques for expert systems in Processing of Complex Ores (New York: Pergamon) 529–536 Hodouin D and Flament F 1989 Frames for the representation of knowledge in Processing of Complex Ores (New York: Pergamon) 553–564 Tardieu H, Rochfeld A and Colletti 1998 La methode MERISE. Tome 1: Principes et outils (Paris: Editions d’Organisation) Mihailescu M, Samuila A, Urs A, Morar R, Iuga A and Dascalescu L 2002 IEEE Trans, Ind. AppLM 1174–1181.

© 2004 by Taylor & Francis Group, LLC

361

Comparative analysis of computer-simulated and experimental sparking voltage of the wire-plate system Z Dudzicz Department of Electrical Engineering and Automatic Control, Technical University of Opole. 45–233 Opole, Sosnkowskiego 31. Poland Abstract. The long term purpose of this paper is to develop new design for zones of charging and depositing of dust particles in electrostatic precipitators and dust measuring instruments. For this purpose, experimental voltage-current (U-I) characteristics of a basic (highvoltage) wire-plate system are first constructed. Particular attention is paid to analysis of sparking voltage in the wire-plate system. Results of measurements are computer processed and analyzed by means of the Matlab package. On the other hand, computer simulations for selected UI characteristics are developed and their results compared to the experimental ones.

1. Introduction A wire-plate system is the fundamental construction element of almost all types of electrostatic precipitators and other facilities operating under high-voltage corona discharge, such as electrostatic separators, band-pass filters and feeders of charged dust particles. Experimental analyses are very time-consuming (and expensive), hence application of computer simulations to determine selected sparking voltage and U-I characteristics can be of great importance when designing electrofiltration facilities. Effective combination of experimental studies and simulation analysis of electroprecipitation phenomena enables the development of new filter designs. 2. Experimental analysis of sparking voltage versus electrode distance for a wireplate system Analysis of sparking voltage, initial voltage of corona discharge and U-I characteristics aims at testing the electrical strength of interelectrode space as well as the dedusting efficiency of electrostatic precipitators. A wire-plate system is used in small two-zone dust collectors, where it operates as a charging zone in which the particle charging process occurs. It thus affects dedusting efficiency. The experimental analysis is made for the following wire diameters (in mm): 0.2, 0.55, 0.8, 1.2, 1.5, 2.0 and 3.0, with wire length of 0.58 m. An aluminium plate shaped as a 500 mm diameter disc is used as a plane electrode. A measuring part is protected by a protection ring of 551–571 mm, in diameter which plays the role of the Rogowski profile. A measuring disc along with the ring is fastened onto the

© 2004 by Taylor & Francis Group, LLC

362

Figure. 1. Sparking voltage of arrangement wire-plate as a function a wire diameter.

insulator plate. Both the disc plate and the discharge electrodes are fixed onto insulation supports in a way that ensures a continuous control of the inter-electrode distance. The following wires are used as discharge electrodes. 1. 2. 3.

Kanthal wire with diameter of 0.2 mm, used in photocopying printer KS-2. Steel wire with of 0.55, 0.8, 1.2, 1.5 mm with chemical composition; 99 % Fe, 0.1 % Cu, 0.3 % Mn, 0.1 % Ni. Wire made of brass as rod brass with diameter of 3.0, 4.0 mm.

Voltage was measured by electrostatic kilovoltmeter. Current values were measured by magnetoelectric microampermeter LG-1. The measuring system was supplied with high DC with negative polarity by means of ABK-70 cable power supply apparatus. All measurements were recalculated to normal conditions 20° C and 760 mm Hg. Fig 1 shows the results of empirical measurements of sparking voltage as a function of electrode distance. As the distance increases, the value of sparking voltage, Up is also growing. This curve may be approximated by straight lines. One curve contains up to 8 measuring points. Each point on the characteristics represents an average value from five experimental measurements. The average value differs insignificantly from individual measuring values. These differences reached a maximum of 3–4 % of the average value. As an example two random cases are given below:

© 2004 by Taylor & Francis Group, LLC

363 20.5 average, measured values 20.2, 19.8, 19.6, 21.2, 21.5 17.0 average, measured values 16.3, 16.5, 17.0, 17.5, 17.5

Figure 2. Sparking voltage of wire-plate set-up as a function of wire diameter.

A dashed line with angle inclination of 45 degree represents a constant value of electric field intensity, equal to 10 kV/cm. It is an average value measured for a non-uniform electric field. This value allows us to estimate the electric strength of this arrangement of electrodes. The curves Up=f(h) can be approximated by means of a straight line. For distances below h=25 mm the route curves are increased slightly. Next follows a flex point, which in the next curves moved towards lower values. At h>25 mm, this route becomes linear. The angle of inclination of the curves is smaller by about 10 degrees than the angle of the broken line (dashed line), that corresponds to the average value of electric field strength. The increase in h value from 5 to 35 mm causes an increase value of sparking voltage by about 20.12 kV. All curves represent similar routes and change directions at the same points. Distances between curves are to a large degree constant, only upper sections of these characteristics are more diversified. Along with the increase of h for each curve, the average value of Up also increases. Each Up=f(h) curve representing a wire of larger diameter is placed higher than the previous one. 3. Sparking voltage experimental research of a wire-plate arrangement as function of wire discharge electrode diameter Analysis of the influence of the change in sparking voltage as a function of the wire discharge electrode diameter appears very interesting. The intention has been to determine the mode in which the electric strength of the interelectrode zone changes. The characteristics

© 2004 by Taylor & Francis Group, LLC

364 performed on the basis of experimental measurements are shown in Fig. 2. All characteristics have very similar routes. Along the first section, the curve represents increased Up values similar to saturation curve. Within the diameter range of 1.5, 2.0 mm an inflexion point occurs and from this point onwards the route is similar to a straight line. These curves compared to those in Fig. 1 are evenly drawn on the picture. The curves are placed more evenly with nearly equal distances between each other. When the distance is constant the value of sparking voltage, Up increases proportionally with increasing wire electrode diameter. This increase in Up which occurs with increase in wire diameters from 2.0 to 4.0 mm amounts to the following values of h; 5.5, 9.4. 10.5, 19.1, 10.9, 10.8 kV. The average values for the electric field intensity for the same curves are; 19.0, 9.4, 7.0, 5.1, 4.4, 3.6 kV/cm. A relationship between the sparking voltage and wire diameter has been observed, however the increase is a minor one. Inclination of a single curve was restricted to the limits of 15–20 degrees. A more significant increase has been observed for curves of bigger distances h. The highest value of electric field intensity growth Ep, occurs for wire kanthal of 0.2 mm. Next, the Ep value strongly decreased below the average value which amounted to 7.35 kV/cm. Decrease in sparking voltage Ep is compared best to the decrease in E value. 4. Computer simulation Computer simulation used for this paper concerned values which were calculated from empirical measurements and theoretical computer simulation on the basis of partial measurements only. In this case the computer modelling and simulation relate closely to partial selected results of measurements. Pure theoretical simulation of phenomena of electric partial discharges and correlation to measurement of sparking voltage or voltage-current characteristics value is very risky. Such processes occur inside non-uniform electric field and additionally run in the polyphase environment. The computer simulation is a tool to process experimental measurement results. On the basis of partial measurements only, we can simulate sparking voltage, voltage-current characteristics saving time and labour otherwise required for full scale experimental measurements. When partial measurements are completed, we know Up=f(h) at d=constant and Up=f(d) at h=constant, and the angle of inclination, and are able to employ computer simulation for further curves for other h and d values. Such accurate approximation of curves is enough for technological purposes. Another solution is to apply partial approximation, that is to split a curve into sections and to approximate subsequent sections separately [3]. The area on Fig.2 can be divided into two different regions. In the first one the saturation curve increases gradually and the second is similar to a straight line. For these regions different methods of calculation of Up=f(d) were applied. It was assumed that the curves were distributed uniformly in the region between the highest and the lowest curve. Comparison of experimental results to theoretical calculated values proves consistency at 10% accuracy. Such accuracy can be accepted only for evaluation of probable sparking voltage level in industrial applications.

© 2004 by Taylor & Francis Group, LLC

365 Sparking voltage as a function of electrode distance can be approximated by straight lines at 5–10% accuracy. 5. Comparison of the research results with Lewitow’s investigations. As indicated by Lewitow’s research, with the same arrangement of wire-plate, the whole range of wire electrode diameters can be divided into two separate areas. In the first area the sparking voltage does not depend on the wire diameter of discharge electrode. The average electric field intensity, at the moment a breakdown spark occurs, is always constant and equals 20 kV/cm. Thus, sparking voltage Up changes proportionally to electrode distance [1]. Up= Ep (h- r)

(1)

This area is called the constant region gradient. The constant region gradient occurs only for wire diameters d< 2–3 mm. This area is wider for greater distances of electrodes, h . The research was carried out for h within the range of 20–100 mm, and wire diameter of 0.8 mm to 10 mm. In the second part of the wives for d> 3mm, the sparking value, Up decreased. On the basis of the experimental measurements a generalised curve has been obtained. Up/Uo—f(h/r)

(2)

This area has been called the similarity region. For similarity region an empirical equation has been found, to calculate the value of Up sparking voltage [1]. Up/Uo=(0.306 h/r—0.642) ln (h/r)

(3)

When the diameter increases above 10 mm, another renewed increase in sparking voltage may be expected as the distribution of the electric field became slightly non-uniform. The critical values of diameter, where the diameter moves from the constant gradient region to the similarity region, may be obtained by employing the aforementioned equations [1,2]. Comparison of the author’s and Lewitow’s measurements points to significant differences. Constant gradient region for wire diameter below 3 mm has not been achieved. Our investigations were carried out for smaller range of distances h=4–40 mm. While Lewitow’s distances range was h=20–200 mm. Our investigations were carried more accurately. The value of strength intensity is constant but by two levels lower. The Up relationship increased accordingly. No sparking voltage decrease has been noted in any other similarity region. 6. Conclusions As the initial value of corona discharge depends on wire diameter, therefore the initial value of electric field intensity, Eo and sparking Ep have to depend on discharge electrode wire diameter. As indicated by the author’s earlier research, values of field intensity, both initial Eo and spark Up, reached very high values within the range of a few hundred kV/cm.

© 2004 by Taylor & Francis Group, LLC

366 The most interesting processes of corona breakdown take place at the surface of discharge electrode of small diameter. The process of corona breakdown depends strongly on wire diameter and the voltage value. For very small distances, corona discharge does not occur and the value of the breakdown sparking voltage is equal to the initial voltage of this arrangement. The intensity of corona discharge and breakdown increased together with the decrease in size o f diameter of the discharge electrode. At very thin wires, the value of strength intensity is so high that it disturbs strongly the average value of the electric field intensity [3, 4]. Research carried out and the discussion described above point to the following conclusion: 1. 2. 3.

Characteristics of sparking voltage and voltage-current can be successfully investigated employing computer-aided experimental research. Within the determined regions 5
The investigations were supported with funds of the Scientific Research Committee within the research project No. 4 T10A 031 22. References [1] [2] [3] [4]

Osnowy elektrogazodinamiki dispiersnych sistem Energia Moskwa 1974 Dymowyje elektrofiltry. Energia. Moskva 1980 Dudzicz Z 2002 Elektrofiltry 2002. Kraków. Komputerowa symulacja rozkladu pola elektrycznego przestrzeni mi´dzyelektrodowej elektrofiltru. p.78–84 Bachtajew Sz A Koronnyj razrjad na mikroprowodach, Nauka Kazachskoj SSR, 1984

© 2004 by Taylor & Francis Group, LLC

367

Decomposition of diesel particulate materials and nitric oxides using a dielectric barrier discharge Yukihiko Yamagata, Takafumi Matsui, Takashi Ebihara, Katsueori Muraoka Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, Kasuga Fukuoka 816–8580, Japan Abstract. We have demonstrated the usefulness of a trapping-filter trapping-low temperature burning scheme for diesel participate materials (DPM) aftertreatment in an actual diesel exhaust gas using a dielectric barrier discharge (DBD). The DPM trapped by a mesh electrode were decomposed to CO and CO2 molecules using the DBD. Simultaneous decomposition of the DPM and nitric oxides in diesel exhaust gas is also presented.

1. Introduction Emissions of diesel participate materials (DPM), nitric oxides (NOx) and hydrocarbons (HC) from diesel engines and other combustion processes are becoming more and more strictly regulated from health concern. Therefore, there have been many attempts to reduce these emissions using diesel participate filters, catalysts, absorption/thermal-desorption or by combinations of these effects [1–5]. However, because of lack of sufficient action and/or excessive cost, promising techniques are yet to be explored. Recently, thermodynamic non-equilibrium gas discharges, such as packed-bed discharge [6], corona discharge [7] and dielectric barrier discharge (DBD) [2–5,8], have attracted much attention for these purposes. For example, packed-bed discharges have been recognized to be able to decompose NOx and volatile organic compounds (VOC). However, it is not suitable for use in high gas flow rate due to large pressure loss caused by the structure. Moreover, most of the electrical energy is consumed for excitation and ionization of predominantly existing nitrogen and oxygen molecules instead of the decomposition of these materials, because these hazardous materials are exhausted in very small concentrations of a few tens to hundreds ppm. Therefore, the decomposition efficiency using thermodynamic non-equilibrium gas discharges for these hazardous materials is very low. In order to extract the potential of these discharges for decomposition, densification or localization of these hazardous materials is prerequisite. We have proposed a new decomposition technique for environmentally hazardous materials with low concentration [3,4,9], and achieved successful decomposition of VOC [9]. This technique is based on the combination of a DBD with densification/localization and a trapping-filter or honeycomb-structured adsorbents. This combination does not spoil the high gas flow rate, and the DBD can easily generate non-equilibrium plasmas on a large discharge space. We consider that this concept is the most suitable for the decomposition of hazardous materials in very small concentrations. In this paper, we demonstrate this technique for the decomposition of DPM and NOx exhausted from an actual diesel engine.

© 2004 by Taylor & Francis Group, LLC

368 2. Experimental setup Figures 1 and 2 show the illustration of the cross section of a discharge reactor and the experimental set up for the decomposition of DPM and NOx in an actual diesel exhaust gas, respectively. The box shape electrode, which one side opened, was made from a stainless steel mesh (mesh size of 2 µm) and placed in a discharge reactor. This mesh electrode acts as a trapping filter of DPM and a grounded electrode. A powered-electrode covered on both sides by a set of mica sheets (thickness of 1 mm) was inserted in the mesh electrode with a separation of 1 mm from the each mesh surface. The DBD was generated between the powered-electrode and the mesh electrode by applying a high voltage (up to 10 kV) with a frequency of 1 to 10 kHz. Figure 3 shows the photograph of the DBD generated at 7 kV (3 kHz). As can be seen from this photograph, the DBD was uniformly generated in the reactor. Part of the exhaust gas from a diesel engine (0.273 l, 3600 rpm, maximum power 6.0 PS) was fed into the discharge reactor. The gas flow rate was changed from 1 to 5 l/min. As the

Fig. 1 Illustration of the cross section of a discharge reactor.

Fig. 2 Experimental set up for the decomposition of DPM and NOx in an actual diesel exhaust gas.

© 2004 by Taylor & Francis Group, LLC

369

Fig. 3 Photograph of the DBD generated at 7 kV (3 kHz).

exhaust gas passed through the mesh electrode as a trapping filter, DPM were trapped near the surface of the mesh electrode. After sufficient accumulation of DPM over the filter, the DBD was generated between the powered-electrode and mesh electrodes to decompose the trapped DPM. Using this configuration, the DPM are localized (condensed) in a limited space, which ensures an effective decomposition of the DPM. The differential pressure across the mesh electrode increases when a large amount of DPM is trapped, while the pressure decreases when the DPM are decomposed by DBD. So the differential pressure is an important parameter to indicate the degree of DPM trapping. Thus, we measured the differential pressure, the gas concentration, gas temperature and the smoke density in order to understand the performance of this reactor. 3. Results and discussion In order to check the reliability of the mesh electrode as a trapping filter for the DPM, we measured the smoke density before and after the mesh electrode. Figure 4 shows that the smoke density changes as a function of load power for the diesel engine. The smoke density before the mesh electrode increases with increase load. As for the 100 % load, the DPM density was estimated to be 98 µg/l. On the other hand, the smoke density after the mesh electrode is nearly zero at any load of the power. This result suggests that the mesh electrode traps almost all of the DPM. As can be seen in Fig. 3, the DBD is uniformly generated between the powered-electrode and the mesh electrode. From these results, this electrode configuration is suitable for the decomposition of the DPM trapped by the mesh electrode. The DBD was periodically generated in order to demonstrate the trapping and the decomposition of the DPM using the discharge reactor. Figure 5 shows the temporal changes

Fig. 4 Smoke density changes as a function of load power for the diesel engine.

© 2004 by Taylor & Francis Group, LLC

370

Fig. 5 Temporal changes of a differential pressure across the mesh electrode, the smoke density before the mesh electrode, and the gas temperature in the discharge reactor.

of a differential pressure across the mesh electrode, the smoke density before the mesh electrode, and the gas temperature in the discharge reactor. Each DBD was sustained at 7 kV with 3 kHz for 10 min, and the load of the diesel engine was 100 %. The differential pressure increased initially with the time passage due to the trapping of the DPM by the mesh electrode. On the other hand, the differential pressure rapidly decreased as soon as the DBD was generated, and was kept at a low value while the DBD was sustained. This feature occurred periodically, as seen with the DBD treatment in Fig. 5. The DPM mainly consists of carbon and HC. As the DPM are exposed to the DBD, the DPM should be decomposed into CO2 and H2O through reactions with surrounding oxidizing atoms and molecules, such as oxygen and NO2. Figures 6 (a) and (b) show the concentrations of CO2 and CO molecules after the discharge reactor, respectively. These concentrations were measured at several times for the non-DBD and subsequently-occurred DBD periods. The CO2 and CO concentrations for the DBD periods are larger than that for non-DBD periods although there are some fluctuations in the concentrations. Furthermore, water drops were observed on the discharge reactor walls after several discharge treatments. This

Fig. 6 Concentrations of CO2 and CO molecules after the discharge reactor.

© 2004 by Taylor & Francis Group, LLC

371 is caused by the DPM decomposition into H2O. These experimental results suggest the DBD can effectively decompose the DPM into CO, CO2 and H2O and regenerate the mesh electrode for the next DPM trapping. Figure 7 shows the temporal change of NQx concentration after the mesh electrode. The DBD was periodically generated in the same experimental conditions as for Fig. 5. As soon as the DBD was generated, the NQx were also decomposed simultaneously by the DBD. After the rapid decrease, the NCbc concentration slightly increased with time. At the same time, the DPM might be already decomposed by the DBD because the differential pressure was low, as seen in Fig. 5. Thus, the NOx concentration change seemed to be interlocked with the DPM decomposition. A nitric monoxide (NO), which is a main content in diesel exhaust gas, is easily oxidized to NO2 in the discharge following reaction (1). Then the DPM is directly oxidized by NO2 following the reaction of (2). NO+O→NO2

(1)

NO2+DPM→CO2+N2

(2)

These results suggest that the combination of the DBD with a densification/localization technique may be able to achieve simultaneous decomposition of DPM and NOx from diesel exhaust gas. In order to exploit a more advanced idea that may revolutionarily change the aftertreatment of diesel exhaust gases, we have now been trying to consider new scheme. Here, NOx and HC are to be adsorbed in adsorbents, similarly as VOC described above, and DPM are collected to the adsorbent surface by the principle of the electrostatic precipitator. Then, a barrier discharge is formed between the parallel electrodes, inducing desorption and destructions of the NOx, HC and DPM simultaneously. Further results will be published soon.

Fig. 7 Temporal change of the NOx concentration behind the mesh electrode.

© 2004 by Taylor & Francis Group, LLC

372 4. Conclusion We have demonstrated a successful decomposition of the DPM and NOx in an actual diesel exhaust gas using the combination of a DBD with densification/localization using a trapping—filter. The DBD can effectively decompose the trapped DPM into CO, CO2 and H2O, and regenerate the mesh electrode for the next DPM trapping. The NOx were simultaneously decomposed by the DBD. These results suggest the usefulness of this system for the diesel exhaust gas aftertreatment. Acknowledgements This work has been partially supported by Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, Culture and Technology of Japan. References [1] [2] [3] [4] [5] [6] [7] [8] [9]

Hawker P, Hüthwohl G, Henn J, Koch W, Lüdders H, Lüers B and Stommel P 1998, Society of Automotive Engineering 980189 374–380. Chae JO, Hwang JW, Jung JY, Han JH, Hwang HJ, Kim S and Demidiouk VI 2001, Phys. Plasmas 8 1403–1410. Muraoka K, Yamagata Y and Ebihara K 2001, Proc. 1st Korea-Japan Joint Symposium on Energy and Environment (KIER, Korea, Dec. 13–14, 2001) 3–4. Yamagata Y, Matsui T, Ebihara T and Muraoka K 2002, Proc. 2nd Japan- Korea Joint Symposium on Energy and Environment (Kyushu Univ., Fukuoka, Japan, Oct. 15, 2002) 29–30. Eliasson B and Kogelschatz V 1991, IEEE Trans. Plasma Sci. 19 309. Yamamoto T, Ramanathan K, Lawless PA, Ensor DS, Newsome JR, Ramsey GH and Plaks N 1992, IEEE Trans. Indust. Applicat. 28 528–534. Kim HH, Takashima K, Katsura S and Mizuno A 2001, J. Phys. D-Appl Phys. 34 604–613. Goto N, Kudo S, Motoyama H and Ohyama S 2002, Jpn. J. Appl. Phys. 41 L64–L66. Urae H, Yamagata Y, Muraoka K, Yamada K, Yamauchi H and H. Okano H 2002, Trans. IEE Japan, 122-A 965–976 (Japanese).

© 2004 by Taylor & Francis Group, LLC

373

Question and Answers

Section 1: Hazards Ignition hazards associated with earthing and bonding of charged conductive objects K Schwenzfeuer and M Glor 19 Question (John Chubb) Have you considered the use of a wood earthing stick as a simple way of removing charge from a conducting body without risk of a spark discharge? Obviously the wood should not be too wet or too dry—although a wet surface will go into a spray discharge before contact and so act like a corona discharger. Butterworth et al reported some studies at an earlier Electronics Conference on the surface resistivity range needed to avoid occurrence of spark type discharges. Author’s Reply (Schwenzfeuer) A piece of wood was not tested, but we did discharge tested with a glove. The surface of this glove was so called “charge dissipative”. What happened was, that a brush discharge occurs during the approach and this discharge was still incendive for ethylene. It must be assumed, that a piece of wood would behave similar and an incendive brush discharge could not be excluded totally. Section 2: MEMS and Applications Microscale electric induction machines for power applications C Liver more, A Forte, T Lyszczarz, S D Umans and J H Lang

45

Question (M Horenstein) How do you deal with the issue of bearing friction? Author’s Reply (Carol Livermore) The rotor is supported both axially and in-plane on gas dynamic bearings. The rotor does not contact its housing; rather, they are always separated by a thin film of pressurized air. There are substantial viscous losses associated with the relative motion of the components when they are separated by micron-scale gaps. Experiments indicate that our Couette flow model is an accurate description of these losses, and we simply fold the losses into our system level models. For relatively larger microgenerators, viscous losses can equal 50% of the input mechanical power, which is undesirable.

© 2004 by Taylor & Francis Group, LLC

374 Question (Keith Davies) The small gaps preclude gas electrical breakdown but do you feel that field emission might eventually limit the possible electric stress? Author’s Reply (C Livermore et al) I am not certain what the ultimate limit will be. I have tested stators for damage under high voltage conditions by applying a phase-to-phase ac voltage while inspecting the electrode structure under a microscope. (The gaps between neighbouring electrodes are 3 ˜m—µ4 m). If there is a defect in the lithography, such as an asperity, breakdown occurs and electrode material is visibly removed. For a small number of defects, the breakdown is a local event that does not significantly impair stator functioning. More widespread breakdown, such as would occur with rough polysilicon electrodes, does impair functioning and leads to a detectable increase in current. In the absence of widespread initial flaws, we have maintained 300V phase-to-phase at MHz frequencies without observable damage or current rise. We have also applied up to 150V peak-to-peak voltage to the stator electrodes with a 2 mm gap between stator and rotor. Upon disassembly no stator damage was observed. Question (Edward Law) As you move from mm-size gas to µm-size gaps, does the classical Newtonian concept of gas viscosity hold strictly true? Are further refinements beyond the usual Cunningham correction factor required? Author’s Reply I have not considered further refinements; you raise a point that is worth looking into. I have not to date noticed a discrepancy between observed viscous torque constant and that which is expected from a Couette flow with the classical gas viscosity. Micro-machined variable capacitors for power generation P Miao, A S Holmes, E M Yeatman, T C Green and P D Mitcheson

53

Question (Lucian Dascalescu) Have you been able to evaluate to what extent gravity affects the operation of your device? In other words is the device able to operate equally well in horizontal and vertical position? Author’s Reply (P Miao) We have tested the device at both positions. We did not see any significant difference. This is because the gravitational force is relatively small compared to the acceleration force of the vibration table. The mass is very light ~ 4 g).

© 2004 by Taylor & Francis Group, LLC

375 Electrostatic charging of trigger actuated spray devices L F Gaunt and J F Hughes

59

Question (Muntomaa) How did you measure the change of the spray, and were all the droplets collected. Author’s Reply (L F Gaunt) We collected the spray droplets in a Faraday Pail to measure their charge. As the spray plume is not very divergent, all droplets were collected. Question (Keith Davies) Ejection of a charged spray leaves a counter charged in the can. What happens to it? Author’s Reply (L F Gaunt) The counter charge accumulating in the liquid reservoir is grounded through the user. This can be achieved by co-extension of a conducting strip in the plastic bottle. Question (J M Chubb) What advantages does your tribo charging offer compared to piezo electric units that are widely used for gas igniters? Author’s Reply (L F Gaunt) Our technique offers just one solution and others may also be conceived. However, cost is an important consideration, and the cheapest technique must be pursued. Question (Prof Bacachandran) How did you determine the threshold Qfm value? Did you calculate the electric field strength at the atomization region? What is the typical value of this field? Author’s Reply (L F Gaunt, J F Hughes) The threshold 2/m value was found experimentally, using the wrap-around phenomenon as an indicator of electrostatic forces coming into play. As the droplet size distribution is wide, this represented a threshold value relating to the specific spray. During the design stage for the novel bigger spray, the electric field strength at he atomization region will be modelled in order to maximise it.

© 2004 by Taylor & Francis Group, LLC

376 Electrostatic forces on ion-charged toner particles D A Hays and J Q Feng

77

Question (Keith Davies) Might toner rounding due to wear in a fluidized bed cause the adhesion of the ion charged toner to be less than the adhesion of toner ion charged in the alternating electric field? Author’s reply (D A Hays) One can only speculate why the toner adhesion measurements reported by Christy from the Moore Research Center at Grand Island, NY were lower than expectations. It doesn’t seem possible that toner would be rounded since the interparticle forces in fluidized beds are actually quite gentle. It seems more likely that since the toner was simultaneously charged and deposited onto a roll, most of the ion charging occurred after the toner deposition. If the ion charging was predominantly on the caps (top) of the toner, the electrostatic adhesion will be much less than expectations for uniformly charged toner. Section 3: Bioelectrostatics Bio-nanotechnology of DNA based on electrostatic manipulation M Washizu

89

Question (David Bakewell) What was the conductivity of the solution when restriction enzymes and RNA polymerase were used in your experiments? Author’s Reply (M Washizu) We cannot stretch DNA in buffer solutions with high conductivity. We use de-ionized water for stretching, and after immobilizing DNA, we replace the solution to the buffer that is optimum for RNA synthesis. Question (Gary Stevens) Do you use particles smaller than 1mm, or AFM stylus, or even carbon nano-tube to immobilize enzymes and do DNA surgery? Author’s Reply (M Washizu) 1) If you use small particles in laser manipulation, many particles will be trapped and coagulated at the focal spot. We use particles whose diameter is about equal to the beam spot size. 2) Electrostatic DNA stretching is done in a thin layer of water sandwiched by a microscope slide and a cover slip. It is inaccessible by mechanical stylus. Of course you can remove the cover slip, but DNA may be broken by hydrodynamic force, and drying takes place very quickly.

© 2004 by Taylor & Francis Group, LLC

377 Cell sorting and separation using dielectrophoresis D Holmes and H Morgan

107

Question (M Horenstein) How do you know the polarization versus frequency for the individual cells so that you can choose the proper operating frequencies? Author’s Reply (D Holmes) Generally, for the experiments shown here we have used all types of known dielectric properties. Either measured by electro rotation in our lab or values taken from the literature. However, one could imagine using a mixture of particles, some known (control particles) and other particles of unknown properties. The relative position of the particles on the separation could then be used to predict the properties of the unknown cells. Question You showed separation of two particle sites (3 µm &. 0.914 µm). What is the expected resolution for particle separation based on size difference) using your system, and also what is the minimum particle site you can manipulate using this technique? Author’s Reply (D Holmes) Particles as small as DNA macromolecules have been manipulated using this system. Work is being undertaken on particle size separation by altering switching times. Section 4: Measurements A wide bandwidth probe for electrostatic discharge measurements J M Smallwood and G L Hearn

125

Question (Mark Zaretsky) Have you optimized the shield and tip geometry for creating and monitoring a discharge? Author’s Reply (J M Smallwood) The probe has a 20 µm hemispherical tip to the shield, because this is a commonly used probe size in conventional spherical ESD probes. The ESD probe central conductor protrudes slightly from the shield hemisphere profile. This is to increase the probability that the ESF current occurs to the probe conductor rather than the shield. Of course there is a trade-off here. If the tip protrudes too much then at high fields corona discharge may occur or several ESD smaller events may occur instead of one larger event. The protrusion must be kept small to reduce this effect.

© 2004 by Taylor & Francis Group, LLC

378 Contact charging method for the measurement of charge decay in electrostatic dissipative materials J Paasi, T Kalliohaka, T Luoma, R Ilmén and S Nurmi 137 Question (John Chubb) 1) Have you made, do you plan to make, any comparison of your contact charging method to tribo charging? 2) I would also express my concern about the validity of contact charging as a generally applicable method. Comments are included on our Website (www.sci.co.uk) comparing the merits and limitations of various methods of charge decay measurement. A major problem with contact charging is that relatively insulating components on the surface may not be effectively charged. The response to earthing the sample shows mainly the fastest decay route within the material, not at the surface. Author’s Reply (Jaakko Paasi) 1) We have done some comparative tests but good comparison is difficult because of low reproducibility of tribocharging. 2) I fully agree with Dr. Chubb that the contact charging method fails in fully characterising the charge decay behaviour in some materials but the situation is similar with all existing methods. We can always find such materials which will not be fully characterised by using a single method. Therefore, complementary measurements using two or three different methods should be recommended for complex materials. A particle charge spectrometer for determining the charge and of individual dust grains on Mars S Fuerstenau and G Wilson 143 Question (John Chubb) To concentrate your particle flow along the axis of your sampling tube, I would commend use of a clean air sheath with a streamline flow. We did this in the monitor for airborne fibres we developed, and reported at an earlier Electrostatics Conference. I wondered about using time varying electric fields to get aerodynamic/inertia size information for particle sizing. This might be a longitudinal field or better a lateral field with examination of the relative signals seen on side vs other side sensing. This would be more compatible with the time response needed. Author’s Reply (Fuerstenau) We are planning to use a clean air concentric flow in order to confine particles to the center streamlines of the flow. Your sheath flow design may be very appropriate. The inertia and drag forces on the relatively large particles are large compared to the electric forces. Measuring variations in the particle velocity dye to the imposition of an electric field will probably be problematic.

© 2004 by Taylor & Francis Group, LLC

379 Question (R G Harrison) The triboelectric dust charging leads to a negatively skewed particle charge distribution. Will additional sources of ionisation, e.g. cosmic rays, lead to a more symmetrical charge distribution on Mars? Author’s Reply (S Fuerstenau et al) Cosmic rays in Mars’ atmosphere will generate small ions, of both polarities. These will discharge particles tending to push the particle charge distribution towards zero and therefore towards the normal symmetrical charge distribution for the particles. However, to the extent that dust is a source of net charge in the atmosphere, the introduction of charge pairs from cosmic ray ionization is not expected to change the balance of charge in atmosphere on the whole. Cosmic ray ionization will shift the charge distribution on the particles but will not remove any net charge asymmetry due to the introduction of triboelectrically charged particles. Measurement of optical intensity and fluence generated by spark discharges J C Crager and M N Horenstein

149

Question (W D Greason) The current probe measures all of the discharge current but it is not known what percentage of the light output is captured by the photodetector. Your comments would be appreciated. Author’s Reply (Crager & Horenstein) It is true that the current probe measures the total current and the photodetector measures only a portion of the light emitted. Since it is difficult to capture all of the light, we are assuming here that the light has a constant radiation pattern. Therefore, we have been measuring the relative light emission from the spark gap under the assumption that a change in the current would be accompanied by a relative charge in the emitted light. Question (Ulrich von Pidoll) I think the quality of the capacitor used in the discharge circuit is very important concerning the energy actually dissipated in the discharge. This effect is the greater, the greater the capacitance and the greater its inductance. Have you made a test series with a nearly ideal capacitor (air capacitor, pulse capacitor for MHz)? Author’s Reply (Joseph C Crager and Mark Horenstein) No we have not performed tests with nearly ideal capacitors. We are, however, currently running experiments with smaller capacitors (in the 100pF range). This is a good suggestion and we will take it into consideration.

© 2004 by Taylor & Francis Group, LLC

380 Section 6: ESD Electrostatic testing of ESD-protective clothing for electronics industry J Paasi, S Nurmi, T KaUiohaka, G Coletti, F Guastavino, L Fast, A Nilsson, P Lemaire, J Laperre, C Vogel, J Haase, T Peltoniemi, G Reina, A Borjesson and J Smallwood 239 Question (John Chubb) Would you not expect that when a garment surface is rubbed the charge will spread quickly over a large area of the garment and will then have a large area for the charge to dissipate through the undergarments to the earthed body underneath? In this case surely seam-seam conductivity bonding is not too critical? Author’s Reply (Paasi et al) Charge can dissipate from the garment through normal clothing to the earthed body of the operator under proper conditions (good contacts between the ESD garment normal clothing and the operator’s body; sufficient conductivity of the normal clothing materials is also required. In a general situation, such a dissipation path is not guaranteed. Therefore, good conductivity over the garment seams is a necessity. Question (Keith Davies) Why is the MBM damage threshold set at 2KV when real body voltages can be greater than 10kV? Author’s Reply (J Smallwood) I’m not aware of how this target level was determined. I understand that this is a take-off between the economic benefit of providing more rugged devices and thus preventing excessive ESD losses, and the additional cost and possible performance losses associated with providing a higher level of ESD protection. Question (N L Allen) You referred to hazards in terms of voltage, but surely the important criteria are charge and stored energy? Author’s Reply (J Paasi) I used voltage just because it is common practice of the field (ESO risks in electronics industry). Clearly, voltage is a poor guide in defining risks, because for a given charge, voltage depends on the capacitance of the object (which can vary more than an order of magnitude). Charge and energy would be much better guides for risk assessment. Even more better would be charge and peak ESO current, as will be shown in the following talk/ paper by Smallwood and Paasi.

© 2004 by Taylor & Francis Group, LLC

381 Section 7: Environment Electrostatics and the environment G S P Castle

309

Question (Mark Horenstein) With regard to lightning you pointed out that the earth’s “leaky capacitor” would quickly discharge were it not for the charge regeneration mechanism. If that regeneration mechanism begins with the negative side of the polarized cloud discharging to ground, is there any speculation as to how the remaining positive charge migrates to the ionosphere? One would think that the positive charge would migrate in the fairweather field down to earth. Author’s Reply (G S P Castle) The positive charge in the upper part of the cloud does not migrate as High as the ionosphere. It is driven upwards in the cloud by thermodynamically induced updrafts. At a height of about 50km, the air is conductive enough that it essentially an equipotential. This positive charge produces the “fairweather field” of about 100V/m directed downwards and yes, left alone it would cause positive charge to migrate to earth and discharge the earth in a matter of minutes. It is the localized regeneration due to thousands of nearly simultaneous lightning strikes across the globe that replenishes the negative charge on the earth and the positive charge in this layer. Question (John Chubb) What is the mechanism by which lightning discharges get started at the top end to get the leader started? I understand that discharges start with emission of VHP that suggest discharges between individual charged droplets may be the initiating mechanism. Author’s Reply (G S P Castle) I have no special insights into the microscopic details of this initiation. However it seems reasonable that localized field perturbations leading to microdischarges between charged droplets could initiate the stepped leader. Comment (J Smallwood) A few years ago I was involved in an EC funded project that developed belt-based electrostatic separators and produced similar results to yours, from post consumer plastic waste streams. My view is that the main problems for viable recycling are that; 1) virgin material is so cheap that it is difficult to have a viable recovery process without subsidy, 2) the market distrusts recycled material and prefers to buy virgin material, so we have a problem of market education. Author’s Reply (G S P Castle) I agree with both your observations, we have experienced similar reluctance on the part of plastics users for the reasons you cite. However I believe that both concerns are becoming

© 2004 by Taylor & Francis Group, LLC

382 less of a constraint. Certainly in the long term virgin plastic prices can only increase and the viability becomes more obvious. However, even at current prices profitability has been demonstrated because of the low energy consumption of the process and the high purities and recoveries that are now possible. In addition, legislative encouragement to recycling is becoming more common in jurisdictions throughout the world.

© 2004 by Taylor & Francis Group, LLC