- Email: [email protected]
Microwave-induced combustion of volatile organic compounds in an industrial flue gas over the magnetite fixed-bed
Available online at www.sciencedirect.com
Chemosphere 69 (2007) 1821–1826 www.elsevier.com/locate/chemosphere
Technical Note
Microwave-induced combustion of volatile organic compounds in an industrial flue gas over the magnetite fixed-bed Bing-Nan Lee b
a,*
, Wai-Tsen Ying b, Yao-Ting Shen
a
a Department of Leisure and Recreation Management, Kao-Fong College, Pingtung, Taiwan, ROC Division of General Education, Yung Ta Institute of Technology and Commerce, Pingtung, Taiwan, ROC
Received 6 November 2006; received in revised form 10 May 2007; accepted 13 May 2007 Available online 4 September 2007
Abstract A modified domestic microwave oven was applied to heat a magnetite (Fe3O4) fixed-bed for continuous decomposition of volatile organic compounds (VOCs), such as acetone, n-hexane, and dichloromethane (DCM), in a simulated flue gas which contains VOCs equivalent to 2000 ppmv as DCM. Experimental results revealed that effect of the addition of water to the inlet stream on decomposition of DCM in the overall experiment was insignificant. Bulk temperature of the Fe3O4 fixed-bed was also found to reach 600 C from an initial room temperature by 6.5 min under microwave radiation, even though the inlet gas was at a high gas hourly space velocity of 5240 h1 and a high relative humidity of 75%. Moreover, the VOCs in the inlet stream could be decomposed completely over the Fe3O4 fixed-bed by microwave heating at a power level of 645 W at heating time of 10 min. The conversion of VOCs is stable when the Fe3O4 fixed-bed has been heated longer than 10 min with microwave radiation. The microwave-induced heating upon Fe3O4 fixed-bed processing appears to be not only an energy efficient technique for air pollutions treatment but also a promising technology for variety of VOCs in a flue gas from industrial factory being decomposed simultaneously and completely. 2007 Elsevier Ltd. All rights reserved. Keywords: Microwave radiation; Acetone; n-Hexane; Dichloromethane
1. Introduction Volatile organic compounds (VOCs) have been used as extracting solvents, degreasers, cleaning agents, and lubricants in various industrial processes resulting in emission in the atmosphere and health risk to human (Jennings et al., 2001). Stringent regulations to control the VOC emissions have been implemented in countries for growing environmental concerns. For example, the European Community has set stage emissions limit to 35 g of total organic compounds (TOC) per m3 gasoline loaded. Similarly, an emission limit of 10 g TOC m3 has been established by
*
Corresponding author. Tel.: +886 8 7624002x462; fax: +886 8 7622397. E-mail address: [email protected] (B.-N. Lee). 0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.05.030
the United States Environmental Protection Agency Standard 40 CFR Part 63 (Khan and Ghoshal, 2000). A variety of VOC control technologies including thermal oxidation, catalytic incineration, photo degradation, and photocatalytic oxidation etc., have been applied to decompose VOCs to final products such as H2O and CO2. Thermal oxidation systems usually combust VOCs at relatively high temperatures (700–1000 C); however, low inlet concentrations will require greater heat input (Neyestanaki et al., 1995; Ismagilov and Kerzhentsev, 1999; Sinquin et al., 2000; Chen et al., 2002; Wang et al., 2003). Although catalytic treatment processes can be carried out at mild temperatures, deactivation of catalyst may occur at low-temperature (T < 350–400 C) and suffer from chloride ion (Cl) poisoning (Lago et al., 1996). Furthermore, an incomplete combustion may result in the rise of highly toxic by-products. For example, Au/Co3O4 catalyst can convert dichloromethane (CH2Cl2) into CO2
1822
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
and HCl at temperatures higher than 350 C; however, CH2Cl2 tend to form chloroform and carbon tetrachloride, which are more toxic than CH2Cl2, at temperature below 250 C (Chen et al., 1996). Also, the utilized catalytic materials require proper disposal as a hazardous waste if they are not recyclable (Khan and Ghoshal, 2000). Compared with conventional heating techniques, microwave heating offers not only higher heating rate but also material selective heating. Chen et al. (1984) revealed that dark colored compounds can be heated rapidly with microwave radiation to high temperature (>1000 C). There is a tremendous thermal gradient from the interior of the char particle to its cool surface due to the fact that microwave heating is both internal and volumetric heating (Guo and Lua, 2000). Walkiewicz et al. (1988) also showed that microwave energy can be effective in the heating of magnetite (Fe3O4) to a maximum temperature of 1258 C within 2.75 min. Further works found that microwaves could be used to enhance solvent extraction of contaminants from soil (Punt et al., 1999) as well as removal of VOCs from polluted air (Ai et al., 2005). Uslu and Atalay (2003) conducted an experiment of microwave-assisted desulfurization of coal in the presence of magnetite, and higher temperatures were observed in the vicinity of magnetite. Therefore, decomposition of the VOCs by microwave radiation heating upon a Fe3O4 fixed-bed processing seems to be an attractive alternative. In the present work, acetone (ACE), n-hexane (n-HEX), and dichloromethane (DCM) are used as VOCs to investigate the ability of a microwave-assisted Fe3O4 fixed-bed process for air pollutants treatment. In addition, the effect of process parameters including microwave power and heating time, and relative humidity (RH, %) and gas hourly space velocity (GHSV, h1) of the inlet gas were also examined.
2. Experimental 2.1. Materials Analytical grade ACE, n-HEX, and DCM were purchased from Merck Co. (Whitehouse station, NJ, USA) and used without any further purification. The commercial magnetite used as microwave radiation absorber was obtained from Osaka Co. (Kogyo, Japan) in form of pellets between 3 and 6 mm. Prior to the experiments being carried out, Fe3O4 pellets were dehydrated in an oven at 105 C for two days and stored in a desiccator before used. Air flow was generated from an air compressor (AC-152, YuhSheng Co., Taiwan) and the air was filtered with a filter set (SPF-080, Shin-Shin Inc., Taiwan) to remove particulates and oil droplets before entering into the experimental apparatus. 2.2. Experimental setup A domestic microwave oven (NE-R30A, International Inc., Taiwan) with a maximum power of 750 W and frequency of 2450 MHz, was modified as a continuously variable power setting mode. Fig. 1 shows the microwave setup (as named CMWR) which was applied to carry out all the experiments. Magnetite pellet was weighted (ca. 500 g) and filled in a reactor made of quartz (50 mm i.d. and 40 cm length) as a fixed-bed in the CMWR system. The air stream was set at various desired flow rates by a mass flowmeter (FQM-10, Shin-Shin Inc., Taiwan). Part of the purified air flowed through a VOC solvent-containing bottle and then merged with the main air stream to obtain desired concentrations of VOC. The humidity of the inlet stream was measured with a humidity meter (TH-380, Li-Shen Inc., Taiwan) and was maintained at a desired constant level (RH = 75 ± 2% @ 25 C) during
Fig. 1. Experimental setup of the microwave-assisted magnetite fixed-bed. Legend: 1. air compressor; 2. filter set; 3. mass flowmeter; 4. check valve; 5. VOCs solvent-containing bottle; 6. water bottle; 7. Fe3O4 fixed-bed; 8. microwave oven; 9. contol box; 10. temperature recorder; 11. K-type thermometer; 12. needle valve, 13. CO2 and CO analyzer.
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
the experiment runs by adjusting the flowrate of carrier gas into a deionized water bottle prior to being introduced down-flow to the reactor. Bulk temperature of the Fe3O4 fixed-bed during microwave treatment was regularly measured with four K-type thermometers at different locations in depth of the central positions in the fixed-bed. The temperatures were recorded simultaneously by an electronic control system with a computer program (DATA-168, Double-flow Tech., Taiwan). The air stream containing of ACE, n-HEX, and DCM was also sampling in a 500 ml flexible Tedlar bag (SKC 232-02, SKC Inc., Eighty Four, PA, USA) periodically during microwave treatment at both inlet and outlet of the reactor for the expected VOC qualitative and/or quantitative analysis. The inner pressure of the quartz reactor was kept at 1.05 atm and the GHSV was calculated with respect to the volume of the Fe3O4 fixed-bed and flow rate of the inlet stream. For instance, an inlet stream with a flow rate of 428.5 ml s1 is calculated at a GHSV of 5240 h1. After concentration of VOCs in the inlet stream was equivalent to that of the outlet stream, microwave radiation was activated and uniformly followed for 40 min. Besides, concentrations of CO2 and CO in the effluent stream were determined throughout the experiment with a CO2 and CO analyzer (EGA-200, Shanq-Tzuoo Tech., Taiwan). 2.3. Chromatographic analysis The collected gas samples were analysed with a gas chromatograph equipped with a flamed ionization detector (HP 6890, Hewlett-Packard Co., Alpharetta, GA, USA). A HP5 capillary column for 30 m in length with stationary phase thickness of 0.25 lm and a diameter of 0.32 mm (J&W Inc., USA) was utilized for high-resolution chromatographic separation. Nitrogen gas (purity = 99.999 vol%) was used as both carrier gas and make-up gas at a flow rate of 4 and 20 ml min1, respectively. Column head pressure of nitrogen gas was set at 0.82 kg cm2. A typical experiment the GC oven temperature was programmed as follows: 50 C initial for 1 min, then ramp of 22.5 C min1 up to 95 C and hold for 1 min. Injector port temperature was held at 200 C and detector temperature maintained constant at 250 C. Calibration was performed with external standards.
1823
is the height of the reactor in the microwave inner cavity and is ca. 22.5 cm. As shown in Fig. 2, the higher the microwave power output was, the higher temperature of the bed could reach. This result is similar to that found in the research by Kingman et al. (2004). Guo and Lua (2000) indicated that an electric power (P) dissipated within the material is proportional to the square of the electric field strength (E) and the voltage frequency (f) under a uniform microwave field. It can be express as P ¼ e0 e00 E2max pf , where e0 represents dielectric constant of free space (8.85 · 1012) and e00 is the real part of the complex permittivity. Assume that no heat transfer occurs, the increase in temperature (DT) as a function of the specific heat (c) and the mass (m) of the material can be expressed as DT ¼ e0 e00 E2max pft=cm or DT = Pt/cm, where t is the time of exposure to microwave energy. This suggests that an increase of E will give rise to more rapid heating on the Fe3O4 fixed-bed. In this study, a reduction in the required heating time was observed when a higher power level was applied. For instance, a shorter time period (in 8 min) was observed for the Fe3O4 fixed-bed being heating up to 910 C from room temperature (25–28 C) at a power level of 645 W while 14 min was required as 572 W was applied. In addition, the Fe3O4 fixed-bed during microwave heating at a power level of 645 W or 572 W showed an initial increase in temperature followed by a slow down and finally reach a constant. This is attributed to microwave penetration decreases as a result of temperature of the material rises. Ramesh et al. (1999) indicated that the total power absorbed by the material mainly depends on the reflection coefficient (p) and attenuation constant of the material. For a homogeneous material, p can be presented as a function of er , which is the root-mean-square value of the complex dielectric constant. Attenuation constant, however, is a product of the free space permittivity and the relative dielectric constant of the material. At a fixed microwave frequency, both the attenuation constant and the loss tangent of the material, which determines the total power dissipated within the sample, depend solely on 1050 900
3.1. Effect of microwave power on heating a magnetite fixed-bed
Temp. (ºC)
750
3. Results and discussion
600 450 300
The magnetite pellets were heated under microwave radiation at four power levels and each run under the absent of inlet gas was held for 30 min. Bulk temperature of the Fe3O4 fixed-bed was measured at the location with a dimensionless depth ratio (Zi/L) of 0.9, where Zi represents distance from the location in the reactor paralleled to the ceil of the microwave inner cavity. In addition, L
150
361 W
525 W
572 W
645 W
0 0
300
600
900
1200
1500
1800
Microwave radiation time (s) Fig. 2. Effect of microwave power on heating of the magnetite fixed-bed.
1824
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
temperature. In contrast, a slow heating rate was obtained while microwave power was set at 525 W or below, and a two-step temperature increment of the Fe3O4 fixed-bed was observed. This implies that a longer microwave heating time interval is required prior to oxidation of VOCs in proceeding when a lower power level is applied during the microwave heating process. 3.2. Effect of relative humidity of the inlet gas on decomposition of DCM Sinquin et al. (2000) showed that the total decomposition of CCl4 needs a higher temperature in the absence of water in comparison with tests carried out in the presence of water on both the LaCoO3 and the LaMnO3 + d catalysts. On the contrary, several researches indicated that the presence of water vapor in the feed stream may lead to deactivation of catalyst in a catalytic oxidation of VOC process (Lago et al., 1996; Lou and Lee, 1997). Nevertheless, water has been well known to be the most effective microwave energy absorber. In order to investigate whether in the presence of water in the inlet gas leads to a negative effect on conversion of VOCs under microwave radiation. The inlet streams over the Fe3O4 fixed-bed under the microwave heating experiments were conducted either in the absence of water (RH < 3%) or in a high level of water vapor (RH = 75 ± 2%). The inlet stream was also prepared containing 1000 ppmv of DCM and was set at a GHSV of 1572 h1. As shown in Fig. 3, the time period required for halftransformation of DCM in the presence of water was seen approximately closed to that of the experiment conducted in the absence of water. This may be due to the phenomenon that only little of microwave energy was absorbed by the water vapor in the inlet stream and resulted in a little decrease in heating rate of the Fe3O4 fixed-bed during the experiment proceeding. Kinston and Jassie (1988) revealed that the temperature of material is raised, the dissipation
factor decreases. This decrease occurs because the dielectric relaxation time of water increases while the water temperature increases. Therefore, the rotational frequency of the water is further out of coincidence with the input microwave angular frequency, and then absorption decreases. Moreover, bulk temperature of the Fe3O4 fixed-bed either in the presence or the absence of water experiment was measured above 556 C, which is the ignition temperature of DCM, under microwave radiation within 6 min. Complete conversion of DCM was also found both in the presence and the absence of water at microwave radiation time of 10 min. It can be concluded that effect of the addition of water to the inlet stream on decomposition of DCM in the overall experiments was insignificant. This is due to the fact that Fe3O4 does not play a role of catalyst but is as a thermal medium in combustion of DCM process under microwave radiation. Therefore, a relative humidity of 75 ± 2% in the inlet stream was chosen to perform the following experiments. 3.3. Effect of gas hourly space velocity of the inlet stream on decomposition of DCM The influence of GHSV on the conversion of DCM over a magnetite fixed-bed under microwave radiation is shown in Fig. 4. The inlet stream was prepared containing 2000 ± 25 ppmv of DCM at room temperature in order to simulate the industrial flue gas. On the basis of the experiment data, there was a little difference of DCM conversion in the microwave heating process for the inlet streams at GHSV of either 1572 h1 or 5240 h1. Bulk temperature of the Fe3O4 fixed-bed was also found to reach 600 C from an initial room temperature by 6.5 min under microwave radiation, even though the inlet gas was at a high GHSV of 5240 h1. Consequently, the conversion of DCM at this time period was calculated close to 50%. This outcome suggests that internal temperature of the Fe3O4 pellets may be much higher than that of its surface as a result of the internal and volumetric nature of microwave heating. Therefore,
1050
1
100
0. 9
1050
900 900
0. 8
RH= 75
0. 4
450
RH < 3
0. 3
RH= 75
300
750
1572 h-1 -1
5240 h
60
600
-1 ⎯ ⎯ 1572 h ⎯ 5240 h-1
40
450
Temp.(ºC)
600
RH < 3 0. 5
Conversion (%)
0. 6
Temp. (ºC)
Ci /C 0
80
750
0. 7
300 20
0. 2
150
150
0. 1 0
0 0
2
4
6
8
10
12
14
16
Microwave radiation time (min) Fig. 3. Effect of relative humidity of the inlet gas on oxidation of dichloromethane under microwave radiation (GHSVinlet gas = 1572 h1, PMW= 645 W, C0, DCM = 1000 ± 15 ppmv).
0
0 0
4
8
12
16
20
24
28
32
Microwave radiation time (min) Fig. 4. Comparison of dichloromethane conversion at different gas hourly space velocity of inlet stream under microwave radiation (RHinlet gas = 75 ± 2%, PMW = 645 W, C0, DCM = 1000 ± 15 ppmv).
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
the microwave-induced chemical reaction can proceed more quickly and effectively at a lower bulk temperature in comparison with conventional surface heating with an electric oven during a catalytic oxidation process. 3.4. Decomposition of acetone, n-hexane and dichloromethane Combustion of VOCs in the inlet stream was carried out at steady state conditions using the Fe3O4 fixed-bed and microwave heating process. All the VOC solvents in the containing bottle were supplied with equal ratio by volume and mixed completely. The total VOCs concentrations in the inlet streams were prepared equivalent to 2000 ± 25 ppmv as DCM, and were adjusted with a RH of 75 ± 2% and at a GHSV of 5240 h1 for all the experiments. Also, the DCM stream was treated without Fe3O4 fixed-bed to investigate if decomposition of VOC was simply due to radiation heat treatment. Results revealed that less than 2% of DCM in the inlet stream was reduced throughout the experiment, which had been carried out for 40 min simply with microwave radiation. The conversion levels of several VOCs streams over the Fe3O4 fixed-bed as a function of microwave radiation time are shown in Fig. 5. In the low-temperature range, such as microwave heating time less than 6 min, conversion levels of VOCs in the streams decrease in the order: mixture of DCM and ACE > DCM only > mixture of DCM and nHEX > mixture of DCM, ACE and n-HEX. The data in Fig. 5 show that bulk temperature of the Fe3O4 fixed-bed at Zi/L of 0.9 under microwave heating was also in the same order of the conversion levels. It appeared that the addition of n-HEX in the inlet streams would reduce the conversion level of VOCs in the low-temperature range of the microwave heating process. This is due to the fact that n-HEX has the lowest ignition temperature (225 C) of the VOCs used in this study and the combustion of nHEX could generate final products such as H2O and CO2, which may absorb part of energy during the microwave heating process. It was also seen that bulk tempera1
1000
0.9
900
0.8
800 700
DCM
C i /C0
0.6
600
0.5
DCM+ACE
500
0.4
DCM+n-HEX
400
0.3
300
DCM+ACE+n-HEX
0.2
200
0.1
100
0
Temp. (ºC)
0.7
1825
ture of the Fe3O4 fixed-bed in the experiment performed with an addition of n-HEX was lowered ca. 89 ± 47 C at the time period 3–7 min in comparison with the run with DCM only. The CO2 concentration in the effluent stream was increased as function of microwave radiation time, up to a maximum value of 1980 ppmv near 10 min and then varied between 1940 and 1970 ppmv for the inlet stream simply contained DCM. It is speculated that the DCM in the inlet stream could completely be decomposed into final products while the CO concentration was measured below 30 ppmv during the microwave heating process. The mass balance of C between input DCM and output CO2 is calculated above 99.9% during the microwave-induced decomposition of DCM process. However, an increase of CO2 concentration ca. 200 ± 20 ppmv was observed during the low-temperature range of the experiment with the inlet stream containing both n-HEX and DCM. In contrast, the deviation of the VOCs conversion levels was observed insignificantly as the microwave heating process was lasting longer than 10 min. At this time period, bulk temperature of the Fe3O4 fixed-bed was measured higher than 900 C and the concentration of the parent VOCs in the outlet stream from the reactor was found to be below the detection limited level by a GC analysis. This implies that complete decomposition of VOCs in the inlet streams over the Fe3O4 fixed-bed can be achieved when the microwave-induced heating process has been prolonged for 10 min. 4. Conclusions The experimental results in this study showed that the addition of water to the inlet stream on decomposition of DCM in the overall experiments was insignificant. This is due to Fe3O4 does not play a role of catalyst but is as a thermal medium in combustion of DCM process under microwave radiation. Decomposition of ACE, n-HEX, and DCM over the Fe3O4 fixed-bed by means of microwave-induced heating process can be achieved completely within 10 min as the inlet streams at a high GHSV of 5240 h1 and RH of 75 ± 2% with a concentration of VOCs equivalent to 2000 ppmv as DCM. The microwave-induced chemical reaction can proceed more quickly and effectively at a lower bulk temperature in comparison with conventional surface heating with an electric oven during a catalytic oxidation process. Therefore, the microwave-induced heating on Fe3O4 fixed-bed processing appears to be both an energy efficient technique and a promising technology for decomposition of a variety of VOCs in a flue gas from industrial factory simultaneously.
0 0
5
10
15
20
25
30
35
40
Acknowledgement
Microwave radiation time (min)
Fig. 5. Conversion of VOCs over the magnetite fixed-bed under microwave radiation (GHSVinlet gas = 5420 h1, RHinlet gas = 75 ± 2%, PMW = 645 W, C0, VOCs = 2000 ± 25 ppmv as DCM).
The authors would like to thank the financial support from the National Science Council, Taiwan, Republic of China (NSC-92-2626-E -127-001).
1826
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
References Ai, Z.H., Yang, P., Lu, X.H., 2005. Degradation of 4-chlorophenol by microwave irradiation enhanced advanced oxidation processes. Chemosphere 60, 824–827. Chen, B., Bai, C., Cook, R., Wright, J., Wang, C., 1996. Gold/cobalt oxide catalysts for oxidative destruction of dichloromethane. Catal. Today 30, 15–20. Chen, F.Y., Pehkonen, S.O., Ray, Madhumita B., 2002. Kinetics and mechanisms of UV-photodegradation of chlorinated organics in the gas phase. Water Res. 36, 4203–4214. Chen, T.T., Dutrizac, J.E., Haque, K.E., Wyslouzil, W., Kashyap, S., 1984. Relative transparency of minerals to microwave radiation. Canadian Metall. Quart. 23, 349–351. Guo, J., Lua, A.C., 2000. Preparation of activated carbons from oil-palmstone chars by microwave-induced carbon dioxide activation. Carbon 38, 1985–1993. Ismagilov, Z.R., Kerzhentsev, M.A., 1999. Fluidized bed catalytic combustion. Catal. Today 47, 339–346. Jennings, V.L.K., Rayner-Brandes, M.H., Bird, D.J., 2001. Assessing chemical toxicity with the bioluminescent photobacterium (Vibrio Fisheri): A comparison of three commercial systems. Water Res. 35, 3448–3456. Khan, F.I., Ghoshal, A.K., 2000. Removal of volatile organic compounds from polluted air – Review. J. Loss Prevent. Proc. 13, 527–545. Kingman, S.W., Jackson, K., Bradshaw, S.M., Rowsaon, N.A., Greenwood, R., 2004. An investigation into the influence of microwave treatment on mineral ore comminution. Powder Tech. 146, 176–184.
Kinston, H.M., Jassie, L.B., 1988. Introduction to Microwave Sample Preparation: Theory and Practice. American Chemical Society, Washington, DC. Lago, R.M., Green, M.L.H., Tsang, S.C., Odlyha, M., 1996. Catalytic decomposition of chlorinated organics in air by copper chloride based catalysts. Appl. Catal. B – Environ. 8, 107–121. Lou, J.C., Lee, S.S., 1997. Destruction of trichloromethane with catalytic oxidation. Appl. Catal. B – Environ. 12, 111–123. Neyestanaki, A.K., Kumar, N., Lindfors, L.E., 1995. Catalytic combustion of propane and natural gas over Cu and Pd modified ZSM Zeolite catalysts. Appl. Catal. B – Environ. 7, 95–111. Punt, M.M., Raghavan, G.S.V., Belanger, J.M.R., Pare, J.R.J., 1999. Microwave-assisted process for the extraction of contaminants from soil. J. Soil Contam. 8, 577–592. Ramesh, P.D., Brandon, D., Scha¨chter, L., 1999. Use of partially oxidized SiC particle bed for microwave sintering of low loss ceramics. Mater. Sci. Eng. A 266, 211–220. Sinquin, G., Petit, C., Libs, S., Hindermann, J.P., Kiennemann, A., 2000. Catalytic destruction of chlorinated C1 volatile organic compounds (CVOCs) reactivity, oxidation and hydrolysis mechanisms. Appl. Catal. B – Environ. 27, 105–115. ¨ ., 2003. Microwave heating of coal for enhanced Uslu, T., Atalay, U magnetic removal of pyrite. Fuel Process. Technol. 85, 21–29. Walkiewicz, J.W., Kazonich, G., MaGill, S.L., 1988. Microwave heating characteristics of selected minerals and compounds. Miner. Metall. Proc. 5, 39–42. Wang, W., Chiang, L.W., Ku, Y., 2003. Decomposition of benzene in air streams by UV/TiO2 process. J. Hazard. Mater. B101, 133–146.
Chemosphere 69 (2007) 1821–1826 www.elsevier.com/locate/chemosphere
Technical Note
Microwave-induced combustion of volatile organic compounds in an industrial flue gas over the magnetite fixed-bed Bing-Nan Lee b
a,*
, Wai-Tsen Ying b, Yao-Ting Shen
a
a Department of Leisure and Recreation Management, Kao-Fong College, Pingtung, Taiwan, ROC Division of General Education, Yung Ta Institute of Technology and Commerce, Pingtung, Taiwan, ROC
Received 6 November 2006; received in revised form 10 May 2007; accepted 13 May 2007 Available online 4 September 2007
Abstract A modified domestic microwave oven was applied to heat a magnetite (Fe3O4) fixed-bed for continuous decomposition of volatile organic compounds (VOCs), such as acetone, n-hexane, and dichloromethane (DCM), in a simulated flue gas which contains VOCs equivalent to 2000 ppmv as DCM. Experimental results revealed that effect of the addition of water to the inlet stream on decomposition of DCM in the overall experiment was insignificant. Bulk temperature of the Fe3O4 fixed-bed was also found to reach 600 C from an initial room temperature by 6.5 min under microwave radiation, even though the inlet gas was at a high gas hourly space velocity of 5240 h1 and a high relative humidity of 75%. Moreover, the VOCs in the inlet stream could be decomposed completely over the Fe3O4 fixed-bed by microwave heating at a power level of 645 W at heating time of 10 min. The conversion of VOCs is stable when the Fe3O4 fixed-bed has been heated longer than 10 min with microwave radiation. The microwave-induced heating upon Fe3O4 fixed-bed processing appears to be not only an energy efficient technique for air pollutions treatment but also a promising technology for variety of VOCs in a flue gas from industrial factory being decomposed simultaneously and completely. 2007 Elsevier Ltd. All rights reserved. Keywords: Microwave radiation; Acetone; n-Hexane; Dichloromethane
1. Introduction Volatile organic compounds (VOCs) have been used as extracting solvents, degreasers, cleaning agents, and lubricants in various industrial processes resulting in emission in the atmosphere and health risk to human (Jennings et al., 2001). Stringent regulations to control the VOC emissions have been implemented in countries for growing environmental concerns. For example, the European Community has set stage emissions limit to 35 g of total organic compounds (TOC) per m3 gasoline loaded. Similarly, an emission limit of 10 g TOC m3 has been established by
*
Corresponding author. Tel.: +886 8 7624002x462; fax: +886 8 7622397. E-mail address: [email protected] (B.-N. Lee). 0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.05.030
the United States Environmental Protection Agency Standard 40 CFR Part 63 (Khan and Ghoshal, 2000). A variety of VOC control technologies including thermal oxidation, catalytic incineration, photo degradation, and photocatalytic oxidation etc., have been applied to decompose VOCs to final products such as H2O and CO2. Thermal oxidation systems usually combust VOCs at relatively high temperatures (700–1000 C); however, low inlet concentrations will require greater heat input (Neyestanaki et al., 1995; Ismagilov and Kerzhentsev, 1999; Sinquin et al., 2000; Chen et al., 2002; Wang et al., 2003). Although catalytic treatment processes can be carried out at mild temperatures, deactivation of catalyst may occur at low-temperature (T < 350–400 C) and suffer from chloride ion (Cl) poisoning (Lago et al., 1996). Furthermore, an incomplete combustion may result in the rise of highly toxic by-products. For example, Au/Co3O4 catalyst can convert dichloromethane (CH2Cl2) into CO2
1822
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
and HCl at temperatures higher than 350 C; however, CH2Cl2 tend to form chloroform and carbon tetrachloride, which are more toxic than CH2Cl2, at temperature below 250 C (Chen et al., 1996). Also, the utilized catalytic materials require proper disposal as a hazardous waste if they are not recyclable (Khan and Ghoshal, 2000). Compared with conventional heating techniques, microwave heating offers not only higher heating rate but also material selective heating. Chen et al. (1984) revealed that dark colored compounds can be heated rapidly with microwave radiation to high temperature (>1000 C). There is a tremendous thermal gradient from the interior of the char particle to its cool surface due to the fact that microwave heating is both internal and volumetric heating (Guo and Lua, 2000). Walkiewicz et al. (1988) also showed that microwave energy can be effective in the heating of magnetite (Fe3O4) to a maximum temperature of 1258 C within 2.75 min. Further works found that microwaves could be used to enhance solvent extraction of contaminants from soil (Punt et al., 1999) as well as removal of VOCs from polluted air (Ai et al., 2005). Uslu and Atalay (2003) conducted an experiment of microwave-assisted desulfurization of coal in the presence of magnetite, and higher temperatures were observed in the vicinity of magnetite. Therefore, decomposition of the VOCs by microwave radiation heating upon a Fe3O4 fixed-bed processing seems to be an attractive alternative. In the present work, acetone (ACE), n-hexane (n-HEX), and dichloromethane (DCM) are used as VOCs to investigate the ability of a microwave-assisted Fe3O4 fixed-bed process for air pollutants treatment. In addition, the effect of process parameters including microwave power and heating time, and relative humidity (RH, %) and gas hourly space velocity (GHSV, h1) of the inlet gas were also examined.
2. Experimental 2.1. Materials Analytical grade ACE, n-HEX, and DCM were purchased from Merck Co. (Whitehouse station, NJ, USA) and used without any further purification. The commercial magnetite used as microwave radiation absorber was obtained from Osaka Co. (Kogyo, Japan) in form of pellets between 3 and 6 mm. Prior to the experiments being carried out, Fe3O4 pellets were dehydrated in an oven at 105 C for two days and stored in a desiccator before used. Air flow was generated from an air compressor (AC-152, YuhSheng Co., Taiwan) and the air was filtered with a filter set (SPF-080, Shin-Shin Inc., Taiwan) to remove particulates and oil droplets before entering into the experimental apparatus. 2.2. Experimental setup A domestic microwave oven (NE-R30A, International Inc., Taiwan) with a maximum power of 750 W and frequency of 2450 MHz, was modified as a continuously variable power setting mode. Fig. 1 shows the microwave setup (as named CMWR) which was applied to carry out all the experiments. Magnetite pellet was weighted (ca. 500 g) and filled in a reactor made of quartz (50 mm i.d. and 40 cm length) as a fixed-bed in the CMWR system. The air stream was set at various desired flow rates by a mass flowmeter (FQM-10, Shin-Shin Inc., Taiwan). Part of the purified air flowed through a VOC solvent-containing bottle and then merged with the main air stream to obtain desired concentrations of VOC. The humidity of the inlet stream was measured with a humidity meter (TH-380, Li-Shen Inc., Taiwan) and was maintained at a desired constant level (RH = 75 ± 2% @ 25 C) during
Fig. 1. Experimental setup of the microwave-assisted magnetite fixed-bed. Legend: 1. air compressor; 2. filter set; 3. mass flowmeter; 4. check valve; 5. VOCs solvent-containing bottle; 6. water bottle; 7. Fe3O4 fixed-bed; 8. microwave oven; 9. contol box; 10. temperature recorder; 11. K-type thermometer; 12. needle valve, 13. CO2 and CO analyzer.
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
the experiment runs by adjusting the flowrate of carrier gas into a deionized water bottle prior to being introduced down-flow to the reactor. Bulk temperature of the Fe3O4 fixed-bed during microwave treatment was regularly measured with four K-type thermometers at different locations in depth of the central positions in the fixed-bed. The temperatures were recorded simultaneously by an electronic control system with a computer program (DATA-168, Double-flow Tech., Taiwan). The air stream containing of ACE, n-HEX, and DCM was also sampling in a 500 ml flexible Tedlar bag (SKC 232-02, SKC Inc., Eighty Four, PA, USA) periodically during microwave treatment at both inlet and outlet of the reactor for the expected VOC qualitative and/or quantitative analysis. The inner pressure of the quartz reactor was kept at 1.05 atm and the GHSV was calculated with respect to the volume of the Fe3O4 fixed-bed and flow rate of the inlet stream. For instance, an inlet stream with a flow rate of 428.5 ml s1 is calculated at a GHSV of 5240 h1. After concentration of VOCs in the inlet stream was equivalent to that of the outlet stream, microwave radiation was activated and uniformly followed for 40 min. Besides, concentrations of CO2 and CO in the effluent stream were determined throughout the experiment with a CO2 and CO analyzer (EGA-200, Shanq-Tzuoo Tech., Taiwan). 2.3. Chromatographic analysis The collected gas samples were analysed with a gas chromatograph equipped with a flamed ionization detector (HP 6890, Hewlett-Packard Co., Alpharetta, GA, USA). A HP5 capillary column for 30 m in length with stationary phase thickness of 0.25 lm and a diameter of 0.32 mm (J&W Inc., USA) was utilized for high-resolution chromatographic separation. Nitrogen gas (purity = 99.999 vol%) was used as both carrier gas and make-up gas at a flow rate of 4 and 20 ml min1, respectively. Column head pressure of nitrogen gas was set at 0.82 kg cm2. A typical experiment the GC oven temperature was programmed as follows: 50 C initial for 1 min, then ramp of 22.5 C min1 up to 95 C and hold for 1 min. Injector port temperature was held at 200 C and detector temperature maintained constant at 250 C. Calibration was performed with external standards.
1823
is the height of the reactor in the microwave inner cavity and is ca. 22.5 cm. As shown in Fig. 2, the higher the microwave power output was, the higher temperature of the bed could reach. This result is similar to that found in the research by Kingman et al. (2004). Guo and Lua (2000) indicated that an electric power (P) dissipated within the material is proportional to the square of the electric field strength (E) and the voltage frequency (f) under a uniform microwave field. It can be express as P ¼ e0 e00 E2max pf , where e0 represents dielectric constant of free space (8.85 · 1012) and e00 is the real part of the complex permittivity. Assume that no heat transfer occurs, the increase in temperature (DT) as a function of the specific heat (c) and the mass (m) of the material can be expressed as DT ¼ e0 e00 E2max pft=cm or DT = Pt/cm, where t is the time of exposure to microwave energy. This suggests that an increase of E will give rise to more rapid heating on the Fe3O4 fixed-bed. In this study, a reduction in the required heating time was observed when a higher power level was applied. For instance, a shorter time period (in 8 min) was observed for the Fe3O4 fixed-bed being heating up to 910 C from room temperature (25–28 C) at a power level of 645 W while 14 min was required as 572 W was applied. In addition, the Fe3O4 fixed-bed during microwave heating at a power level of 645 W or 572 W showed an initial increase in temperature followed by a slow down and finally reach a constant. This is attributed to microwave penetration decreases as a result of temperature of the material rises. Ramesh et al. (1999) indicated that the total power absorbed by the material mainly depends on the reflection coefficient (p) and attenuation constant of the material. For a homogeneous material, p can be presented as a function of er , which is the root-mean-square value of the complex dielectric constant. Attenuation constant, however, is a product of the free space permittivity and the relative dielectric constant of the material. At a fixed microwave frequency, both the attenuation constant and the loss tangent of the material, which determines the total power dissipated within the sample, depend solely on 1050 900
3.1. Effect of microwave power on heating a magnetite fixed-bed
Temp. (ºC)
750
3. Results and discussion
600 450 300
The magnetite pellets were heated under microwave radiation at four power levels and each run under the absent of inlet gas was held for 30 min. Bulk temperature of the Fe3O4 fixed-bed was measured at the location with a dimensionless depth ratio (Zi/L) of 0.9, where Zi represents distance from the location in the reactor paralleled to the ceil of the microwave inner cavity. In addition, L
150
361 W
525 W
572 W
645 W
0 0
300
600
900
1200
1500
1800
Microwave radiation time (s) Fig. 2. Effect of microwave power on heating of the magnetite fixed-bed.
1824
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
temperature. In contrast, a slow heating rate was obtained while microwave power was set at 525 W or below, and a two-step temperature increment of the Fe3O4 fixed-bed was observed. This implies that a longer microwave heating time interval is required prior to oxidation of VOCs in proceeding when a lower power level is applied during the microwave heating process. 3.2. Effect of relative humidity of the inlet gas on decomposition of DCM Sinquin et al. (2000) showed that the total decomposition of CCl4 needs a higher temperature in the absence of water in comparison with tests carried out in the presence of water on both the LaCoO3 and the LaMnO3 + d catalysts. On the contrary, several researches indicated that the presence of water vapor in the feed stream may lead to deactivation of catalyst in a catalytic oxidation of VOC process (Lago et al., 1996; Lou and Lee, 1997). Nevertheless, water has been well known to be the most effective microwave energy absorber. In order to investigate whether in the presence of water in the inlet gas leads to a negative effect on conversion of VOCs under microwave radiation. The inlet streams over the Fe3O4 fixed-bed under the microwave heating experiments were conducted either in the absence of water (RH < 3%) or in a high level of water vapor (RH = 75 ± 2%). The inlet stream was also prepared containing 1000 ppmv of DCM and was set at a GHSV of 1572 h1. As shown in Fig. 3, the time period required for halftransformation of DCM in the presence of water was seen approximately closed to that of the experiment conducted in the absence of water. This may be due to the phenomenon that only little of microwave energy was absorbed by the water vapor in the inlet stream and resulted in a little decrease in heating rate of the Fe3O4 fixed-bed during the experiment proceeding. Kinston and Jassie (1988) revealed that the temperature of material is raised, the dissipation
factor decreases. This decrease occurs because the dielectric relaxation time of water increases while the water temperature increases. Therefore, the rotational frequency of the water is further out of coincidence with the input microwave angular frequency, and then absorption decreases. Moreover, bulk temperature of the Fe3O4 fixed-bed either in the presence or the absence of water experiment was measured above 556 C, which is the ignition temperature of DCM, under microwave radiation within 6 min. Complete conversion of DCM was also found both in the presence and the absence of water at microwave radiation time of 10 min. It can be concluded that effect of the addition of water to the inlet stream on decomposition of DCM in the overall experiments was insignificant. This is due to the fact that Fe3O4 does not play a role of catalyst but is as a thermal medium in combustion of DCM process under microwave radiation. Therefore, a relative humidity of 75 ± 2% in the inlet stream was chosen to perform the following experiments. 3.3. Effect of gas hourly space velocity of the inlet stream on decomposition of DCM The influence of GHSV on the conversion of DCM over a magnetite fixed-bed under microwave radiation is shown in Fig. 4. The inlet stream was prepared containing 2000 ± 25 ppmv of DCM at room temperature in order to simulate the industrial flue gas. On the basis of the experiment data, there was a little difference of DCM conversion in the microwave heating process for the inlet streams at GHSV of either 1572 h1 or 5240 h1. Bulk temperature of the Fe3O4 fixed-bed was also found to reach 600 C from an initial room temperature by 6.5 min under microwave radiation, even though the inlet gas was at a high GHSV of 5240 h1. Consequently, the conversion of DCM at this time period was calculated close to 50%. This outcome suggests that internal temperature of the Fe3O4 pellets may be much higher than that of its surface as a result of the internal and volumetric nature of microwave heating. Therefore,
1050
1
100
0. 9
1050
900 900
0. 8
RH= 75
0. 4
450
RH < 3
0. 3
RH= 75
300
750
1572 h-1 -1
5240 h
60
600
-1 ⎯ ⎯ 1572 h ⎯ 5240 h-1
40
450
Temp.(ºC)
600
RH < 3 0. 5
Conversion (%)
0. 6
Temp. (ºC)
Ci /C 0
80
750
0. 7
300 20
0. 2
150
150
0. 1 0
0 0
2
4
6
8
10
12
14
16
Microwave radiation time (min) Fig. 3. Effect of relative humidity of the inlet gas on oxidation of dichloromethane under microwave radiation (GHSVinlet gas = 1572 h1, PMW= 645 W, C0, DCM = 1000 ± 15 ppmv).
0
0 0
4
8
12
16
20
24
28
32
Microwave radiation time (min) Fig. 4. Comparison of dichloromethane conversion at different gas hourly space velocity of inlet stream under microwave radiation (RHinlet gas = 75 ± 2%, PMW = 645 W, C0, DCM = 1000 ± 15 ppmv).
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
the microwave-induced chemical reaction can proceed more quickly and effectively at a lower bulk temperature in comparison with conventional surface heating with an electric oven during a catalytic oxidation process. 3.4. Decomposition of acetone, n-hexane and dichloromethane Combustion of VOCs in the inlet stream was carried out at steady state conditions using the Fe3O4 fixed-bed and microwave heating process. All the VOC solvents in the containing bottle were supplied with equal ratio by volume and mixed completely. The total VOCs concentrations in the inlet streams were prepared equivalent to 2000 ± 25 ppmv as DCM, and were adjusted with a RH of 75 ± 2% and at a GHSV of 5240 h1 for all the experiments. Also, the DCM stream was treated without Fe3O4 fixed-bed to investigate if decomposition of VOC was simply due to radiation heat treatment. Results revealed that less than 2% of DCM in the inlet stream was reduced throughout the experiment, which had been carried out for 40 min simply with microwave radiation. The conversion levels of several VOCs streams over the Fe3O4 fixed-bed as a function of microwave radiation time are shown in Fig. 5. In the low-temperature range, such as microwave heating time less than 6 min, conversion levels of VOCs in the streams decrease in the order: mixture of DCM and ACE > DCM only > mixture of DCM and nHEX > mixture of DCM, ACE and n-HEX. The data in Fig. 5 show that bulk temperature of the Fe3O4 fixed-bed at Zi/L of 0.9 under microwave heating was also in the same order of the conversion levels. It appeared that the addition of n-HEX in the inlet streams would reduce the conversion level of VOCs in the low-temperature range of the microwave heating process. This is due to the fact that n-HEX has the lowest ignition temperature (225 C) of the VOCs used in this study and the combustion of nHEX could generate final products such as H2O and CO2, which may absorb part of energy during the microwave heating process. It was also seen that bulk tempera1
1000
0.9
900
0.8
800 700
DCM
C i /C0
0.6
600
0.5
DCM+ACE
500
0.4
DCM+n-HEX
400
0.3
300
DCM+ACE+n-HEX
0.2
200
0.1
100
0
Temp. (ºC)
0.7
1825
ture of the Fe3O4 fixed-bed in the experiment performed with an addition of n-HEX was lowered ca. 89 ± 47 C at the time period 3–7 min in comparison with the run with DCM only. The CO2 concentration in the effluent stream was increased as function of microwave radiation time, up to a maximum value of 1980 ppmv near 10 min and then varied between 1940 and 1970 ppmv for the inlet stream simply contained DCM. It is speculated that the DCM in the inlet stream could completely be decomposed into final products while the CO concentration was measured below 30 ppmv during the microwave heating process. The mass balance of C between input DCM and output CO2 is calculated above 99.9% during the microwave-induced decomposition of DCM process. However, an increase of CO2 concentration ca. 200 ± 20 ppmv was observed during the low-temperature range of the experiment with the inlet stream containing both n-HEX and DCM. In contrast, the deviation of the VOCs conversion levels was observed insignificantly as the microwave heating process was lasting longer than 10 min. At this time period, bulk temperature of the Fe3O4 fixed-bed was measured higher than 900 C and the concentration of the parent VOCs in the outlet stream from the reactor was found to be below the detection limited level by a GC analysis. This implies that complete decomposition of VOCs in the inlet streams over the Fe3O4 fixed-bed can be achieved when the microwave-induced heating process has been prolonged for 10 min. 4. Conclusions The experimental results in this study showed that the addition of water to the inlet stream on decomposition of DCM in the overall experiments was insignificant. This is due to Fe3O4 does not play a role of catalyst but is as a thermal medium in combustion of DCM process under microwave radiation. Decomposition of ACE, n-HEX, and DCM over the Fe3O4 fixed-bed by means of microwave-induced heating process can be achieved completely within 10 min as the inlet streams at a high GHSV of 5240 h1 and RH of 75 ± 2% with a concentration of VOCs equivalent to 2000 ppmv as DCM. The microwave-induced chemical reaction can proceed more quickly and effectively at a lower bulk temperature in comparison with conventional surface heating with an electric oven during a catalytic oxidation process. Therefore, the microwave-induced heating on Fe3O4 fixed-bed processing appears to be both an energy efficient technique and a promising technology for decomposition of a variety of VOCs in a flue gas from industrial factory simultaneously.
0 0
5
10
15
20
25
30
35
40
Acknowledgement
Microwave radiation time (min)
Fig. 5. Conversion of VOCs over the magnetite fixed-bed under microwave radiation (GHSVinlet gas = 5420 h1, RHinlet gas = 75 ± 2%, PMW = 645 W, C0, VOCs = 2000 ± 25 ppmv as DCM).
The authors would like to thank the financial support from the National Science Council, Taiwan, Republic of China (NSC-92-2626-E -127-001).
1826
B.-N. Lee et al. / Chemosphere 69 (2007) 1821–1826
References Ai, Z.H., Yang, P., Lu, X.H., 2005. Degradation of 4-chlorophenol by microwave irradiation enhanced advanced oxidation processes. Chemosphere 60, 824–827. Chen, B., Bai, C., Cook, R., Wright, J., Wang, C., 1996. Gold/cobalt oxide catalysts for oxidative destruction of dichloromethane. Catal. Today 30, 15–20. Chen, F.Y., Pehkonen, S.O., Ray, Madhumita B., 2002. Kinetics and mechanisms of UV-photodegradation of chlorinated organics in the gas phase. Water Res. 36, 4203–4214. Chen, T.T., Dutrizac, J.E., Haque, K.E., Wyslouzil, W., Kashyap, S., 1984. Relative transparency of minerals to microwave radiation. Canadian Metall. Quart. 23, 349–351. Guo, J., Lua, A.C., 2000. Preparation of activated carbons from oil-palmstone chars by microwave-induced carbon dioxide activation. Carbon 38, 1985–1993. Ismagilov, Z.R., Kerzhentsev, M.A., 1999. Fluidized bed catalytic combustion. Catal. Today 47, 339–346. Jennings, V.L.K., Rayner-Brandes, M.H., Bird, D.J., 2001. Assessing chemical toxicity with the bioluminescent photobacterium (Vibrio Fisheri): A comparison of three commercial systems. Water Res. 35, 3448–3456. Khan, F.I., Ghoshal, A.K., 2000. Removal of volatile organic compounds from polluted air – Review. J. Loss Prevent. Proc. 13, 527–545. Kingman, S.W., Jackson, K., Bradshaw, S.M., Rowsaon, N.A., Greenwood, R., 2004. An investigation into the influence of microwave treatment on mineral ore comminution. Powder Tech. 146, 176–184.
Kinston, H.M., Jassie, L.B., 1988. Introduction to Microwave Sample Preparation: Theory and Practice. American Chemical Society, Washington, DC. Lago, R.M., Green, M.L.H., Tsang, S.C., Odlyha, M., 1996. Catalytic decomposition of chlorinated organics in air by copper chloride based catalysts. Appl. Catal. B – Environ. 8, 107–121. Lou, J.C., Lee, S.S., 1997. Destruction of trichloromethane with catalytic oxidation. Appl. Catal. B – Environ. 12, 111–123. Neyestanaki, A.K., Kumar, N., Lindfors, L.E., 1995. Catalytic combustion of propane and natural gas over Cu and Pd modified ZSM Zeolite catalysts. Appl. Catal. B – Environ. 7, 95–111. Punt, M.M., Raghavan, G.S.V., Belanger, J.M.R., Pare, J.R.J., 1999. Microwave-assisted process for the extraction of contaminants from soil. J. Soil Contam. 8, 577–592. Ramesh, P.D., Brandon, D., Scha¨chter, L., 1999. Use of partially oxidized SiC particle bed for microwave sintering of low loss ceramics. Mater. Sci. Eng. A 266, 211–220. Sinquin, G., Petit, C., Libs, S., Hindermann, J.P., Kiennemann, A., 2000. Catalytic destruction of chlorinated C1 volatile organic compounds (CVOCs) reactivity, oxidation and hydrolysis mechanisms. Appl. Catal. B – Environ. 27, 105–115. ¨ ., 2003. Microwave heating of coal for enhanced Uslu, T., Atalay, U magnetic removal of pyrite. Fuel Process. Technol. 85, 21–29. Walkiewicz, J.W., Kazonich, G., MaGill, S.L., 1988. Microwave heating characteristics of selected minerals and compounds. Miner. Metall. Proc. 5, 39–42. Wang, W., Chiang, L.W., Ku, Y., 2003. Decomposition of benzene in air streams by UV/TiO2 process. J. Hazard. Mater. B101, 133–146.