Ozone Catalytic Oxidation of Gaseous Toluene over MnO 2-Based Ozone Decomposition Catalysts Immobilized on a Nonwoven Fabric

Degradation of toluene gas by ozone catalytic oxidation (OZCO) by using a MnO2-based ozone decomposition catalyst (ODC) was investigated to clarify the reactive site of ODC material with O3. An optimum structure for the ODC to remove O3 and toluene were proposed. For honeycomb ODC, toluene degradation by OZCO occurred only around the entrance of the honeycomb ODC, and we expected that a thinner ODC would increase the toluene degradation efficiency. A nonwoven fabric on which ODC was immobilized was developed to decompose O3 and volatile organic compounds simultaneously. The toluene degradation ratio and the mineralization of toluene to CO2 were determined to evaluate the performance of the fabric. Furthermore, the effects of relative humidity and O3 concentration on the decomposition and mineralization ratios were also investigated with or without 254 nm UV irradiation (UV254). The fabric decomposed and mineralized toluene to CO2, even at low O3 concentrations. Although high humidity reduced the degradation ratio of toluene, UV254 irradiation improved the recovery of the degradation ratio and increased the mineralization ratio.


INTRODUCTION
Ozone (O 3 ) is a very strong oxidant and has attracted great interest for degrading organic pollutants, such as dyes (Wu et al., 2008a;Cuiping et al., 2009;Srinivasan et al., 2009) and other organic compounds (Perkowski et al., 1996;Yan et al., 2016), in wastewater, and it has been used in pilot-scale wastewater pre-treatment (Lucasa et al., 2010;Somensi et al., 2010).However, the reaction efficiency for ozonation alone is low (Gong et al., 2008;Cuiping et al., 2009;Rao and Chu, 2009) because the reactivity of O 3 depends on the chemical structure and hydrophilicity or hydrophobicity of pollutants, and the pH of the wastewater (Chang et al., 2012).Therefore, O 3 is generally used in the advanced oxidation process (AOP) involving hydroxyl radicals (OH•) combined with other oxidation techniques, such as UV and hydrogen peroxide, including the peroxone process (Hernandez et al., 2002;Garoma and Gurol, 2004;Lee et al., 2007;Wu et al., 2008b;Cuiping et al., 2009;Wang et al., 2009), and electro-peroxone treatment (Wang et al., 2015;Frangos et al., 2016), zerovalent metals (Chand et al., 2009;Wen et al., 2014;Zhang et al., 2015a;Zhang et al., 2015b), photocatalysis (Domínguez et al., 2005;Giri et al., 2008;Mehrjouei et al., 2015), and ultrasonication (Chand et al., 2009;Joseph et al., 2009).Concentrations of O 3 from several to several hundred parts per million by volume are used for AOP techniques in contamination control of gaseous pollutants and wastewater.Especially, O 3 with UV irradiation is widely used for AOP techniques because the initial cost is lower than for other oxidative degradation techniques except for ozonation alone (Mahamuni and Adewuyi, 2010).On the other hand, it has been reported that O atoms, O 2 , and OH radicals are formed from O 3 on the surface of an MnO 2 -based ozone decomposition catalyst (ODC) as follows (Einaga and Ogata, 2009;Zhao et al., 2012;Huang et al., 2015a).
Here, * indicates the species at the catalytic active sites.These radicals can react with volatile organic compounds (VOCs) on the ODC surface and the VOCs are decomposed and mineralized to CO and CO 2 .These reactions are generally called ozone catalytic oxidation (OZCO) and are used to degrade various VOC pollutants in gaseous and aqueous phases (Villasenor et al., 2002;Zhao et al., 2008;Nawrocki, 2013;Jia et al., 2016).In gaseous phases, a high concentration of O 3 is used, and surplus O 3 is released from the reactor outlet to the ambient air as exhaust O 3 .Therefore, surplus O 3 used for contamination control must be decreased and kept within safe levels because long-term exposure to O 3 is harmful to human health even at low concentrations.The removal of O 3 is proportional to the amount of ODC; thus, granulated or honeycomb ODC is often used in large quantities for the complete removal of O 3 .However, to ensure the effective use of the ODC, the part of the ODC surface to which O 3 and VOCs are adsorbed and where the decomposition process occurs should be determined to clarify the optimum structure and amount of ODC materials to install in reactors.
Furthermore, at high relative humidity (RH) in the gas phase, the removal ratio of VOC gas on the ODC surface by OZCO decreases owing to the inhibition of the adsorption of hydrophobic O 3 and VOC gas by water molecules on the ODC surface (Sekiguchi et al., 2003;Jeong et al., 2005).This problem has been addressed by using ODCs mixed with alumina or zeolite, which are functional adsorbents, to increase the contact between O 3 and the ODC surface (Einaga and Futamura, 2004;Einaga and Futamura, 2006;Einaga et al., 2013;Rezaei et al., 2013;Huang et al., 2015a, b).Using these adsorbents for OZCO allows VOCs to be decomposed and mineralized to CO and CO 2 , even under wet conditions (Einaga and Futamura, 2006;Einaga et al., 2013), but only a small percentage of VOCs reacted.UV irradiation of ODC powder increases the amount of VOCs decomposed CO 2 by OZCO (Sekiguchi et al., 2003;Huang et al., 2016).This is because UV irradiation generates OH radicals effectively from O 3 , which is concentrated around the ODC surface, according to the following equations (Reisz et al., 2003;Cuiping et al., 2009).
OH radical generation from H 2 O 2 depends on the light intensity (Rosenfeldt et al., 2006); however, because ODC usually has a granular or honeycomb structure, it is difficult to irradiate it with UV sufficiently.
In this study, the reactive site of commercial honeycomb ODC was investigated by using a laboratory-scale singleor three-step reactor.Based on these results, a thin, flexible ODC material, fabricated from an ODC immobilized on a nonwoven fabric (nonwoven ODC), was developed and the use of the ODC was evaluated.The performance of the nonwoven ODC, namely the decomposition ratio of O 3 , decomposition ratio of toluene, and the complete mineralization ratio of toluene to CO and CO 2 , were investigated by using the same reactor.Furthermore, the effects of O 3 concentration and RH on the decomposition and mineralization ratios were also investigated with or without UV irradiation.

Experimental Setup and Procedure
Fig. 1 shows the experimental setup for testing the performance of honeycomb ODC, and VOC removal by OZCO with nonwoven ODC.Honeycomb or nonwoven ODC and an UV 254 or UV 254+185 lamp (OFU, Sankyo Denki, 4W) were placed in a 0.9 L Pyrex reactor.The main wavelengths of UV 254 and UV 254+185 were 254 nm and 254 nm with 3% of the output power at 185 nm, respectively.For honeycomb ODC, a single-step reactor with 30-mmthick honeycomb ODC (30 t × 1 s ), or a three-step reactor prepared by joining three reactors each containing 10-mmthick honeycomb ODC (10 t × 3 s ), was used with or without UV 254 and UV 254+185 irradiation.UV 254+185 readily generates O 3 and OH radicals in the gas phase by photochemical reactions as follows (Jeong et al., 2004, Chang et al., 2013).
However, a UV 254 lamp without O 3 generation was selected as a light source in the experiment for OZCO of toluene using nonwoven ODC to evaluate the relationship between toluene decomposition and O 3 concentration.UV 254 irradiation can decompose O 3 to form O radicals, similar to the surface reaction of the ODC.The model VOC gas was 40 ppm toluene for testing the performance of honeycomb ODC and 10 ppm toluene for evaluating toluene removal by OZCO with nonwoven ODC.The toluene concentrations were adjusted by mixing the toluene standard gas with purified air that was synthesized from dry N 2 and O 2 .The RH of purified air was less than 10 ppm.For each experiment, 8 or 38 ppm of O 3 was generated by UV 254+185 (OZU, 4 W, Sankyo Denki, Tokyo, Japan) irradiation and mixed into the gas flow.When a high concentration of O 3 (730 ppm) was supplied to the reactor, O 3 was also generated with an O 3 generator (ED-OG-R3Lt, Ecodesign, Saitama, Japan) by electric discharge.The gas flow rates were 5.0 L min -1 for testing the performance of honeycomb ODC, and 1.0 L min -1 for toluene removal with nonwoven ODC.To control RH, water vapor was obtained by passing dried air through a porous polytetrafluoroethylene tube in ultrapure (Milli-Q) water at room temperature.The reaction temperature was maintained at room temperature (25 ± 1°C).The VOC removal experiment was started after the toluene gas was supplied to the reactor, and all experiments were performed at least three times.

Nonwoven ODC
The ODC was a TiO 2 /SiO 2 honeycomb with MnO 2 as the main active element (TSO, Nippon Syokubai, Tokyo, Japan).It was crushed with an agate pestle and mortar to a particle size of about 1.0-5.0µm, and fixed to the nonwoven fabric fibers (Polyolefin fiber, Japan Vilene, Tokyo, Japan).The nonwoven ODC contained 0.358 g ODC in an area of 0.0072 m 2 (6 × 12 cm) and was 0.37 mm thick (Fig. 1).The surface area of the nonwoven ODC was measured with a BET surface analyzer (Flowsorb III-2305, Micromeritics, GA, USA) (Table 1).For comparison, the results for ODC powder and the TiO 2 powder (Degussa P25, Nippon Aerosil, Tokyo, Japan), widely used for photocatalytic reactions, was also described (Sekiguchi et al., 2003).
The nonwoven ODC has small surface area compared with ODC powder or TiO 2 powder (Table 1).This is because ODC powder is embedded in the polyolefin fiber to immobilize it.Therefore, to use the nonwoven ODC, it is necessary to consider the saturation of VOC gas adsorption on the ODC surface (Rezaei et al., 2013;Huang et al., 2016).However, covering the fiber completely protects it from oxidative destruction by O 3 and active species generated from O 3 so the fabric is durable and reusable.

Analytical Method
The toluene concentration in the effluent gas was measured by gas chromatography with flame ionization detection Mineralization ratio (%) 100 ( ) Here, C 0 is the upstream initial concentration, C t is the downstream concentration after time t, C CO2,t and C CO,t are the CO 2 and CO concentrations at time t, and N C is the carbon number of the supplied organic gas.N C value of toluene gas is 7 (Sekiguchi et al., 2010).

RESULTS AND DISCUSSION
Performance Test of Honeycomb ODC Fig. 2 compares the toluene removal ratio using a singleor three-step reactor with or without UV irradiation for honeycomb ODC.For the single-step reactor, O 3 was completely decomposed and about 70% of the toluene gas was removed.However, the removal ratio of toluene decreased gradually with time.This result indicates that intermediate degradation products accumulated on the honeycomb ODC surface (Qi et al., 2016), and active species generated from O 3 reacted with the degradation products and toluene competitively.However, UV irradiation increased the removal ratio by promoting the mineralization of degradation products (Huang et al., 2016), which reactivated the ODC surface for toluene by self-cleaning.A similar self-cleaning reaction of O 3 under UV 254 irradiation was observed for a TiO 2 surface in our previous study (Jeong et al., 2004).For UV 254+185 irradiation, direct photolysis of toluene gas was also expected near the UV lamp, although the 185 nm irradiation had little effect on the toluene removal rate for short residence times in the reactor.Therefore, the increase in the toluene removal ratio by UV irradiation was caused by increasing the active species generation from O 3 by Eqs. ( 1)-(3) on the honeycomb ODC surface.This indicates that degradation products generated from toluene existed around the entrance of the honeycomb ODC because UV light reached only the entrance.Therefore, toluene gas was adsorbed by the whole honeycomb ODC, but O 3 decomposition, and thus toluene degradation by OZCO, only occurred only around the entrance.Fig. 2. Toluene removal ratio for honeycomb ODC for a single-step (30 t × 1 s ) or three-step (10 t × 3 s ) reactor with or without UV irradiation.Initial concentration of toluene: 40 ppm, initial concentration of ozone: 8 ppm, gas flow rate: 5.0 L min -1 , RH: < 10 ppm.
A confirmation experiment was performed with a threestep reactor (Fig. 1).The honeycomb ODC in each reactor was a third of the thickness of the honeycomb ODC used in the single reactor.In the three-step reactor, O 3 was carried to the second and third reactors without being decomposed fully by the first reactor, although the O 3 concentration decreased gradually.The three-step reactor achieved a high toluene removal ratio, especially under UV irradiation (Fig. 2).These results showed that O 3 decomposition and generation of O radicals from O 3 occurred around the entrance of the honeycomb ODC, and that degradation efficiency could be increased by using thinner ODC materials in several steps.Based on these results, we developed a thinner, flexible nonwoven ODC (0.37 mm thick) (Fig. 1).We investigated the O 3 decomposition and OZCO performance of this nonwoven ODC for toluene degradation and mineralization with or without UV irradiation under dry or wet conditions in a single-step reactor.

Effect of O 3 on Toluene Gas Removal
Fig. 3 shows the time profiles of toluene removal ratio for the nonwoven ODC with or without O 3 .Without O 3 , the removal ratio decreased rapidly owing to the low adsorption on the nonwoven ODC because of the small surface area.Toluene was not decomposed on the ODC surface without O 3 and was saturated immediately.However, when O 3 was supplied to the reactor, a stable removal ratio was obtained for a long time.This result indicates that even a very thin nonwoven ODC can decompose toluene effectively with the active species generated from O 3 on the ODC surface.The gas phase reaction does not contribute to the decomposition because the reaction rate of O 3 with toluene is much slower than that of active species such as OH radicals (Song et al., 2007;Huangand Li, 2011).under dry and wet (RH 70%) conditions with or without UV 254 irradiation.A removal ratio of more than 95% was maintained under dry conditions, even though the removal ratio decreased temporarily in the early stages of the reaction owing to desorption of excess O 3 .Furthermore, a small increase in O 3 decomposition was observed when the nonwoven ODC was irradiated with UV 254 , indicating that O 3 decomposition mainly occurred via the catalytic reaction of O 3 trapped on the ODC surface, not the gas phase reaction owing to the short residence time, even though UV 254 has a wavelength of O 3 degradation (Reisz et al., 2003;Cuiping et al., 2009).However, the removal ratio of O 3 decreased considerably under wet (70% RH) conditions without UV 254 , and the same trend was reported elsewhere (Sekiguchi et al., 2003;Jeong et al., 2005).The adsorption of hydrophobic O 3 and toluene may have been prevented by water molecules covering the ODC surface.
To clarify the O 3 degradation of the nonwoven ODC, the effect of changing the initial O 3 concentration from 38 to 730 ppm on O 3 removal ratio was measured.The O 3 removal ratio after 90 min is shown in Fig. 5.When the initial concentration of O 3 was 38 ppm, a high O 3 removal ratio of more than 95% was obtained (Fig. 4).It was thought that the amount of ODC was sufficient for 38 ppm of O 3 under these flow conditions.When the initial concentration of O 3 was changed to 730 ppm which was excessive concentration, the O 3 removal ratio decreased by 60%, although the removal ratio was stable.The amount of O 3 decomposed at 730 ppm was about 40 ppm which was similar to that at 38 ppm which was almost removed, and the concentration around 40 ppm was equivalent to the maximal decomposition amount under these flow conditions.Thus, for nonwoven ODC, the amount of removal O 3 is constant and does not depend on the O 3 concentration.Consequently, the amount of nonwoven ODC required for the O 3 removal can be determined easily.

Toluene Removal under Various Conditions
Fig. 6 shows time profiles of toluene removal ratios for nonwoven ODC under conditions of 38 ppm O 3 ; 38 ppm O 3 with UV 254 irradiation; 38 ppm O 3 with 70% RH; 38 ppm O 3 with 70% RH and UV 254 irradiation; and 730 ppm O 3 .Although a stable removal ratio was obtained under all conditions, the ratio differed greatly under each set of conditions.
The removal ratio of toluene was similar under dry conditions with or without UV irradiation.Therefore, direct photolysis of toluene did not occur because a wavelength of less than 200 nm is necessary to dehydrogenate and   open the benzene ring (Kislov et al., 2004), and toluene did not react with active species generated from O 3 in the gas phase under UV 254 irradiation owing to the short residence time in the reactor (Wang and Ray, 2000).However, the lowest removal ratio was obtained under wet (70% RH) conditions because water molecules prevented hydrophobic toluene and O 3 from reaching the same site on the ODC surface.However, the removal ratio was stable even under wet conditions; thus, the water molecules were adsorbed on the ODC surface at a fixed rate, and toluene and the active species from O 3 reacted smoothly.The toluene removal ratio increased substantially under UV 254 irradiation and wet conditions.This result indicated that the amount of OH radicals was increased by the generation of OH radicals by UV 254 irradiation from O 3 concentrated near the ODC surface (Eqs.( 5)-( 7)), in addition to the OH radicals generated from O 3 by catalytic reactions (Eqs.( 1)-( 4)) on the ODC.The OH radicals generated by both processes contributed to toluene removal around the ODC surface.Therefore, nonwoven ODC generated active species and removed toluene at the ODC surface with UV 254 irradiation even under wet conditions because the nonwoven ODC was sufficiently thin for the UV 254 light to reach all the catalytic sites.Nonwoven ODC may be a useful material for designing multistep reactors for OZCO owing to its effective use of UV light and its thinness and flexibility.
When a high O 3 concentration (730 ppm) was supplied to the reactor, the toluene removal ratio was constant at about 100%.Because the amount of O 3 that can be decomposed at the catalyst surface is constant, the toluene removal can be explained as follows.First, ozonation in the gas phase may have occurred in a part of toluene, even during short residence times in the reactor owing to the high O 3 concentration, even though the reaction rate of O 3 with toluene is not high (Song et al., 2007;Huang et al., 2011), and secondary active radicals could be generated by this decomposing process to advance the toluene removal.In actual, when 730 ppm of O 3 was supplied to the reactor without ODC, the toluene removal ratio in the gas phase reaction only with O 3 was about 32.6%.Second, mineralization via the degradation of intermediate products generated by toluene decomposition at high O 3 concentrations may have increased.The active species generated from O 3 on the ODC surface may attack toluene without being affected by the degradation products.To clarify the effect of mineralization, its efficiency was evaluated under the same conditions.

Mineralization Ratio
Fig. 7 shows the toluene mineralization ratios under the same conditions as in Fig. 6.Under dry conditions without UV 254 irradiation, the mineralization ratio was about 20%, and it increased slightly to 35% under the same conditions with UV 254 irradiation, although both removal rates were similar to those in Fig. 6.Thus, OH radicals were generated from O 3 on the ODC surface by UV 254 irradiation (Eqs.( 5)-( 7)), and contributed to the mineralization of degradation products formed by toluene decomposition.However, there was little generation of OH radicals under dry conditions, and they were competing reactions between toluene and the degradation products (Einaga and Futamura, 2004;Sekiguchi et al., 2011).Under wet conditions (70% RH), the mineralization ratio increased to 50-60%, and it was much higher under UV 254 irradiation.These results indicate that many OH radicals were generated from O 3 on the ODC surface under wet conditions (Eq. ( 4) without UV irradiation and Eqs. ( 5)-( 7) with UV irradiation), and contributed to the mineralization of degradation products formed by toluene decomposition.
Under wet conditions without UV 254 irradiation, OH radicals mainly affected mineralization; the OH radicals were not consumed by toluene decomposition owing to the low removal rate because water molecules prevented toluene adsorption.UV 254 irradiation under wet conditions increased the removal and mineralization ratios of toluene by the OH radical reaction.Therefore, a sufficient amount of OH radicals was formed under wet conditions by UV 254 irradiation around the ODC surface to react with both toluene and the decomposition intermediates generated from toluene.In our previous study, the reaction rate of OH radicals formed by UV irradiation on the TiO 2 surface with intermediates was faster than that of the supplied VOC gas (Jeong et al., 2005).Therefore, OH radicals generated by UV 254 irradiation affected the mineralization of the decomposition intermediates rather than toluene removal.
When a high concentration of O 3 (730 ppm) was supplied to the reactor, the highest removal and mineralization ratios of toluene were obtained, indicating that a large amount of O radicals formed on the ODC surface via Eqs.( 1)-(3), and they showed sufficient oxidization and mineralization.However, OH radicals may be generated by Eq. ( 4) from trace water adsorbed on the ODC surface.However, at high concentrations, O 3 acts as scavenger of OH radicals (Lee et al., 2007;Cheng et al., 2013), preventing the effective use of OH radicals.Based on these results, a high concentration of O 3 contributed to toluene removal and its mineralization on the nonwoven ODC surface; however, O 3 should be kept below safe levels at the reactor outlet to protect human health.Therefore, combining UV 254 irradiation with a low concentration of O 3 that could be decomposed completely on the ODC surface resulted in effective VOC gas degradation by OZCO on the nonwoven ODC under dry or wet conditions.

CONCLUSIONS
The performance of a nonwoven ODC was investigated, focusing on the decomposition ratios of O 3 and toluene, the complete mineralization of decomposition intermediates, and the effects of RH and UV 254 irradiation.The results of this study can be summarized as follows.
(1) O 3 was decomposed rapidly only around the entrance of the honeycomb ODC.Based on these results, a thin, flexible nonwoven ODC was developed.
(2) The amount of ODC required for the O 3 removal can be determined easily, and toluene was decomposed and mineralized to CO 2 by OZCO, even at low O 3 concentrations.(3) The removal ratio of toluene decreased because water molecules on the ODC surface prevented O 3 and toluene adsorption.(4) UV 254 irradiation of the ODC surface under wet conditions increased the removal and mineralization ratios of toluene by increasing the OH radical reaction.Based on these results, the most effective use of nonwoven ODC for VOC gas degradation by OZCO was to combine UV irradiation with a low concentration of O 3 that can be decomposed completely on the ODC surface under dry or wet conditions.This method is promising as an air purification technique which can remove not only VOC gases but also organic particles simultaneously due to its nonwoven fabric structure.

Fig. 1 .
Fig. 1.Experimental setup for honeycomb ODC performance test and toluene removal by ozone catalytic oxidation with nonwoven ODC in the presence or absence of UV irradiation.

Table 1 .
BET surface area of various catalysts.