Department of Ecological Engineering Toyohashi University of Technology 1-1 Tempaku-cho, Toyohashi, Aichi, 441-8580, Japan Corresponding author, A. Mizuno: e-mail: mizuno@eco.tut.ac.jp

Abstract— Various types of atmospheric pressure discharges have been studied for particle removal, and plasma chemical treatment of exhaust gas, volatile organic compounds(VOCs), etc. Selectivity and energy efficiency of plasma chemical processes can be significantly improved if plasma is combined with a catalyst. Generation of stable discharge plasma has been quite difficult inside narrow capillaries of an automobile catalytic honeycomb that has been commonly used. In this work, novel method has been proposed, connecting packed-bed or barrier discharge in series with the honeycomb capillaries. With a DC or pulse high voltage application across the capillaries, streamers are extended from the packed-bed discharge into the capillaries. With this method, ionization can be made inside fine channels of honeycomb catalyst made of insulating materials. This discharge is designated as “honeycomb discharge”. The honeycomb discharges have been applied for cleaning of diesel exhaust. The oxidation and reduction characteristics of NOX have been measured. In the simulated gas test, honeycomb discharge had the NO oxidation ability. This result was indicating that the honeycomb discharge have possibility of diesel exhaust gas cleaning. Keywords — Electrical discharge, Non-thermal plasma, Honeycomb catalyst, Gas cleaning, Exhaust gas aftertreatment

I. INTRODUCTION Exhaust gas from diesel vehicles has been causing serious

air pollution, especially in urban area, due to emission of particulate matter (PM), nitrogen oxides (NOX) and other toxic gaseous pollutants. [1-2] Diesel engines have higher thermal efficiency than gasoline engines, and they produce less carbon dioxide (CO2) in exhaust gas. Due to those advantages and high endurance, it cannot be replaced at least

in near future. Effective gas treatment technology, therefore, must be developed urgently. Conventional three-way catalysts have been used to remove NOX from gasoline engine exhaust gas. [3] However, it cannot be used for diesel exhaust gas, because diesel exhaust gas has rich oxygen content. In addition, sometimes gas temperature is not high enough to activate catalyst. Recently, various types of selective catalytic reduction (SCR) have been investigated for NOX treatment of diesel engine exhaust gas. In particular, urea SCR system is one of the most promising after treatment technology to eliminate NOX emission from diesel engine, because urea SCR system can be used in high oxygen concentration. At low temperature, however, it is difficult to get catalyst activity and quick ammonia generation.

Recently, various types of atmospheric pressure discharges have been developed for electrostatic processes and for plasma chemical processes, such as removing particles, cleaning exhaust gas and volatile organic compounds (VOCs), etc. [4-5] Those discharges are able to induce various reactions at room temperature. In order to improve selectivity and energy efficiency of plasma chemical processes, combination of plasma and catalyst is effective. For example, catalyst pellets can be used in packed bed discharge to improve removal efficiency of nitrogen oxides and VOCs[6-7]. Honeycomb is a commonly used geometry of catalysts. It has been difficult to generate electrical discharge evenly inside a honeycomb. If discharge is generated inside a honeycomb, larger surface area can be used with lower pressure drop for improved chemical reactions.

There are several important works recently reported [8-9] to achieve larger surface area or volume of electrical discharge. In this study, a method to generate electrical discharge in honeycomb catalyst is proposed, and characteristics of honeycomb discharge are examined. In the NOX treatment test, two types of high voltage (positive pulse, negative DC) are applied to generate honeycomb discharge. NOX removal rate of 40% can be obtained when negative DC high voltage is applied. The results suggest that this method can be used for low temperature NOX reduction system.

Satoshi Sato, Kazunori Takashima and Akira Mizuno Fellow, IEEE

Diesel Exhaust gas treatment by honeycomb discharge

II. INSTRUMENTATION

A. Plasma Reactor Figure 1 illustrates the plasma reactor used in this

study. A packed-bed discharge is used to form the preceding discharge for honeycomb discharge. The γ-Al2O3 pellets (3mm) covered with catalyst (Fe-Zeolaite) are set in a quartz glass tube (inner diameter: 32mm) with a stainless steel rod (diameter: 6mm) at the center as the preceding discharge electrode. The stainless steel rod has a hole to flow the gas. The holes are opened at 5mm from the end. The outer surface of the quartz glass tube is covered with an aluminum sheet of 5 mm length as a ground electrode. These stainless steel rod and aluminum sheet are connected AC high voltage power supply to generate packed-bed discharge for the preceding discharge. The AC high voltage used was 30kVp-p, 1 kHz.

The honeycomb (cordilite) is set on the packed layer of the pellets. The honeycomb supports Fe-Zeolaite catalyst. A stainless steel mesh (16mesh) is placed on the honeycomb as a counter electrode to the packed bed. This mesh is connected to a high voltage power supply (pulse or DC power supply) to form the discharge inside the honeycomb. As the negative DC, -15kV, was applied, and as the pulse, positive +10kV0-p, 1kHz was applied,

In the packed-bed, discharges are generated at contacting points of the pellets. With the DC high voltage application between the honeycomb, inside channels of the honeycomb can be ionized, as if the AC discharge (packed bed discharge) is sliding into the honeycomb by the DC electric field.

Figure 1 Schematic diagram of the reactor with honeycomb

Figure 2 shows photographs of the discharge. The preceding discharge (packed bed discharge) is shown in (a) and the honeycomb discharge is shown in (b). With the AC

and DC application, the honeycomb discharge is taking place, and uniform light emission along the honeycomb can be observed. Intensity of the light emission increased with increasing the DC voltage until sparking.

(a) Preceding discharge with AC only

(packed bed discharge)

(b) Honeycomb discharge with AC and DC

Figure 2 Photographs of the honeycomb discharge

AC : 30kVp-p, 1kHz, DC: -15kV

B. Experimental Setup Figure 3 depicts the Experimental Setup used in this study.

Concentration of N2, NO and NH3 gas was controlled with a mass flow controller. The gas flow rate was 5L/min. Dry Air was used as a carrier gas, and was preheated to about 350℃ by a tube furnace. The gas temperature was set to 130℃. The gas was fed to the reactor from the honeycomb side towards the pellet layer. The plasma reactor was placed in a convection oven in order to control the ambient temperature. Gas was quantitatively analyzed using a FT-IR (BIO-RAD FTS3000) spectrometer.

A DC high voltage power supply (Pulse Electronic Engineering Co., Ltd. HDV-50K3SUD), pulse high voltage power supply (EGC-KOKUSAI PPS8000) and an AC high voltage power supply (TREK Model20/20C High Voltage Amplifier) equipped with a function generator (Agilent Function/Arbitrary Waveform Generator 33220A) were used. Waveform of the applied voltage was measured using

a digital oscilloscope (Tektronix TDS 2014) equipped with a high voltage probe (Tektronix P644A). Figure 4 shows typical waveforms of applied voltage and discharge current.

Figure 3 Experimental Setup The waveform of AC high voltage was measured using a

digital oscilloscope. The waveform of total discharge current was measured with a non-inductive shunt resistor. A 1k-ohm resistor was inserted in the circuit between the grounded electrode and the ground. Figure 4 shows the waveform of AC high voltage and Total current with the DC high voltage was applied.

Figure 4 Waveform of AC high voltage, negative DC high voltage and Total current (Negative DC high voltage: -15kV and AC high voltage: 30kVp-p, 1kHz)

Positive and negative current were observed when the AC

voltage reached the maximum in the positive half cycle a negative current pulse was observed. This result suggests that the negative pulse current can be attributed to the honeycomb discharge.

The waveform of AC high voltage, positive pulse high voltage and total current were measured. Figure 5 shows the waveform of the AC high voltage, the positive pulse high voltage and the total current. The AC high voltage was synchronized with the positive pulse high voltage. Timing of the positive pulse was set to the timing of negative peak of the AC high voltage.

Figure 5 Waveform of AC high voltage, positive pulse high

voltage and Total current (Positive pulse high voltage: 15kV, 1kHz and AC

high voltage: 30kVp-p, 1kHz)

The current peaks of both positive and negative were observed. When the AC voltage reached the maximum in the negative half cycle a positive current pulse was observed. This result suggests that the positive pulse current can be attributed to the honeycomb discharge.

III. RESULTS AND DISCUSSION

A. NOX conversion using negative DC and AC high voltage Figure 6 shows the result of the NOX removal test using

honeycomb discharge with the negative DC high voltage. In this experiment, the preceding discharge power was 2.1W, and that of the honeycomb discharge was 1.5W. Therefore the total discharge power was 3.6W.

When the preceding discharge was turned on, a half amount of NO was oxidized to NO2. In the next step, the honeycomb discharge was generated by applying the negative DC high voltage. The NO concentration was further deceased by 30ppm, and the NO2 concentration increased by 20ppm. Therefore, about 30ppm of NO was additionally oxidized to NO2 by the honeycomb discharge. In this case, the concentration of NOX, however, did not decrease. This result suggests that the honeycomb discharge oxidizes NO to NO2, and that no reductive reaction takes place with the discharge. After turning off the plasma, the

AC

DC

Current

AC Pulse

Current

NO concentration rapidly returned to the initial value.

Figure 6 NOX removal test using AC and negative DC

(Simulated gas 5.0 L/min, NO 200ppm, Air balance,) Figure 7 shows the result of the NOX removal test using

honeycomb discharge by negative DC high voltage with the 200ppm NH3. In this experiment, power of the preceding discharge was 2.1W, and that of the honeycomb discharge was 1.5W. Therefore the total discharge power was 3.6W.

Figure 7 NOX removal test using AC and negative DC (Gas flow rate 5.0 L/min, NO 200ppm, NH3 200ppm, Air balance,)

When the preceding discharge was generated, a half

amount NO was oxidized to NO2. In the next step, the honeycomb discharge was generated by the negative DC high voltage. The NO concentration was deceased further by 50ppm, while the NO2 concentration was not increased, and the concentration of NOX was decreased by 40ppm. This result suggested that NO was oxidized to NO2 and NO2 was reduced by the injected NH3 in the honeycomb discharge.

The selective catalytic reduction occurred on the surface of catalyst in the honeycomb discharge. After the shut off the plasma, NO and NH3 were rapidly returned to the initial value. The plasma reactor was slightly modified to use the layer of pellet catalyst. Figure 8 depicts the modification. Position of the hole in the rod electrode was lowered so that the gas can pass through the layer of the catalyst pellet. Thickness of the catalyst layer was 15mm. The chemical reaction on the pellet catalyst was expected.

Figure 8 Change of the gas flow pass Figure 9 showed the result of the NOX removal test using

honeycomb discharge with the negative DC high voltage. In this experiment, the preceding discharge power was 2.1W, the honeycomb discharge power was 1.5W and the total discharge power was 3.6W.

Figure 9 NOX removal test using the AC and negative DC (Gas flow rate 5.0 L/min, NO 200ppm, NH3 200ppm, Air balance,)

When the preceding discharge was on, the concentration of NO decreased by 100ppm. The concentration of NO2 increased 40ppm. These results suggested the NO was oxidized to NO2 and the generated NO2 was absorbed by the catalyst.

Then, the honeycomb discharge was generated. The NO concentration was further deceased by 20ppm. The NO2 concentration was not increased and the concentration of NOX decreased by 10ppm. This result suggested that the NO was oxidized to NO2 and NO2 was absorbed on the catalyst, and then reduced by the selective catalytic reduction with NH3 on the surface of catalyst. After turning off the plasma, the NO and NH3 were rapidly returned to the initial value.

B. NOX conversion using positive pulse and AC high voltage Figure 10 is the result of the NOX removal test using

honeycomb discharge with positive pulsed high voltage. In this experiment, power of the preceding discharge was 2.1W, and that of the honeycomb discharge was 2.3W. The total discharge power was 4.4W.

Figure 10 NOX removal test using the AC and positive pulse

(Gas flow rate 5.0 L/min, NO 200ppm, Air balance,) When the preceding discharge is applied, a half of NO

(100pm) was oxidized to NO2, or 80ppm of NO2 was generated. In the next step, the honeycomb discharge was generated by the positive pulse high voltage. The NO concentration deceased by 20ppm, and the NO2 concentration increased by 40ppm. 20ppm of NO was oxidized to NO2 and the absorbed NO2 (20ppm) was released from the catalyst using the positive pulse high voltage. However, with the plasma, the concentration of NOX did not decrease in this case. This result suggests that the plasma promoted only oxidation of NO to NO2. After turning off the plasma, NO returned rapidly to the initial value.

Figure 11 shows the result of the NOX removal test using honeycomb discharge with positive pulse, and with addition of 200ppm NH3. In this condition, the power of the preceding discharge was 2.1W, that of the honeycomb discharge was 2.3W. And the total discharge power was

4.4W. When the preceding discharge was applied, 100ppm of

NO was oxidized to NO2, and converted to NO2. 80ppm of converted NO2 remained in the gas flow and 20ppm was absorbed on the catalyst. In the next step, the honeycomb discharge was generated by the positive pulse high voltage synchronizing with the AC high voltage. The NO concentration deceased further by 20ppm, while that of NO2 did not increase. The oxidized NO was reduced with the NH3 on the catalyst. With the plasma, therefore, the concentration of NOX decreased. After the shut off the plasma, NO and NH3 rapidly returned to the initial value. This result suggests that the selective catalytic reduction is taking place in this experimental condition, and that the preceding discharge and the honeycomb discharge is oxidizing NO to NO2.

Figure 11 NOX removal test using AC and positive pulse

(Gas flow rate 5.0 L/min, NO 200ppm, NH3 200ppm, Air balance,)

IV. CONCLUSION Using honeycomb discharge generated by negative DC

high voltage or positive pulse high voltage, characteristics of the promoted plasma chemical reaction has been studied.

The NOX removal rate of the honeycomb discharge with the negative DC high voltage was 65% , and that of the honeycomb discharge by positive pulse was 30% in the conducted experiment. The result indicated that the honeycomb discharge released the absorbed NO2, while with the addition of NH3, selective catalytic reduction took place at low temperature condition of 130 degree C.

REFERENCES [1] E.P.Weijers, A.Y.Khlystov, G.P.A.Kos and

J.W.Erisman, “Variability of particulate matter concentrations along roads and motorways determined

by a moving measurement unit”, Atmospheric Environment 38 (2004 ) 2993-3002

[2] Yasuhiro Daisho, “A Prospect on Automobiles for the 21st Century –Resolving Problems Associated with Environment and Energy”, JSAE Technical Paper, 20024030, pp.18-24, Vol.56, No.1, 2002 (in Japanese)

[3] N. Takahashi, H. Shinjoh, T. Iijima, T. Suzuki, K. Yamazaki, K. Yokota, H. Suzuki, N. Miyoshi, S. Matsumoto, T. Tanizawa, T. Tanakab, S. Tateishi and K. Kasahara, “The new concept 3-way catalyst for automotive lean-burn engine: NOx storage and reduction catalyst”, Catalysis Today, Vol. 27, No. 1-2, pp. 63-69, 1996

[4] Toshiaki Yamamoto, Masaaki Okubo, Tomoyuki Kuroki and Yukio Miyairi, “Nonthermal Plasma Regeneration of Diesel Particulate Filter", SAE Technical paper, SP-1759, No.2003-01-1182

[5] H.J. White, Industrial Electrostatic Precipitation, Addison-Wesley Publishing Co., 1962

[6] Urashima, K. Jen-Shih Chang Jae-Yoon Park Duck-Chool Lee Chakrabarti, A. Ito, T. , “Reduction of NOx from Natural Gas Combustion Flue Gases by Corona Discharge Radical Injection Techniques,” IEEE Transactions on Industry Applications, Vol.34, No.5, Sep/Oct, pp.934-939, 1998

[7] Oda, T. Kato, T. Takahashi, T. Shimizu, K., “Nitric Oxide Decomposition in Air by Using Nonthermal Plasmaprocessing with Additives and Catalyst,” IEEE Transactions on Industry Applications, Vol.34, No.2, Mar/Apr, 1998C.Louste, G.Artana, E.Moreau, G.Touchard, “Sliding discharge in air at atmospheric pressure : electrical properties”, Journal of Electrostatics, Vol 63, 2005, 615-620

[8] S. Sato, K. Hensel, H. Hayashi, K. Takashima, and A. Mizuno, "Honeycomb discharge for diesel exhaust cleaning," Journal of Electrostatics, 2009.