Continuous degradation of BTEX in landfill gas by the UV-Fenton reaction

Abstract

The degradation of benzene, toluene, ethyl benzene and xylene (BTEX) in waste gas using the UV-Fenton reaction has been investigated. By employing a bubble column reactor, a suitable inlet gas flow-rate was found to be important in determining both the handling capacity and the interfacial area between the gas-phase and liquid-phase. Both the large gas–liquid interfacial area and the UV-Fenton reaction in the liquid-phase enhanced the gas–liquid mass transfer process, and thus improved the removal efficiency of BTEX in waste gas. Under the optimized initial conditions (i.e., pH = 3, H₂O₂ dosage = 5.6 mmol L⁻¹, Fe²⁺/H₂O₂ molar ratio = 0.091 and purified air as a carrier gas), dynamic treatment of a gaseous mixture of BTEX was performed and the removal efficiencies of benzene, toluene, ethyl benzene and o-xylene were always maintained above 84%, 92%, 96% and 97%, respectively, during a 240 min reaction time. After the degradation reaction, the removal effect of the UV-Fenton system could be effectively restored by adding 10% of the initial amount of hydrogen peroxide. The obtained results showed that the UV-Fenton process was an effective method for BTEX degradation in continuous waste gas treatment. Compared with other common waste gas treatment methods like physical absorption, photo-catalytic oxidation and biofiltration, the proposed UV-Fenton method had distinct advantages such as in its degradation ability, large handling capacity and low operational cost. Thus, it is considered to be a promising method for waste gas treatment of typical air pollution sources including landfills, livestock farms, pharmaceutical factories and so on.

---

1. Introduction

Landfill gas consists of many air pollutants and it usually causes a serious annoyance in urban areas.¹ As a group of typical pollutants in landfill gas, benzene, toluene, ethyl benzene and xylene (BTEX) have attracted increasing public concerns because of their severe threats to human health (e.g., causing cancer and leukemia).^2,3 A previous study has reported the distinct detection of BTEX in landfill gas and they accounted for a substantial proportion of the total volatile organic compound (VOC) concentration.⁴ Therefore, it is of great significance to develop an effective degradation method for the BTEX in waste gas.

The Fenton process, which is one of the advanced oxidation processes, is an effective degradation technology and has been widely employed in the wastewater treatment field. Due to the highly reactive hydroxyl radicals (with an oxidizing potential as high as 2.8 V) generated in an aqueous solution by Fenton reagents, organic molecules can be fully mineralized to CO₂ and H₂O₂.⁵ Besides, the hydroxyl radical has no selectivity and it can react with most organic compounds and their intermediate products. Up to now, the Fenton process has been extensively explored and many improved techniques including the electro-Fenton, photo-Fenton and Fenton-like systems have been proposed.^6–8 Especially with the synergistic effect of UV light, both the degradation efficiency and the utilization level of hydrogen peroxide are apparently enhanced.^9,10 This is because part of hydrogen peroxide can directly decompose into hydroxyl radicals with the radiation of UV light instead of the general iron-catalyzed decomposition of hydrogen peroxide.¹¹ Furthermore, the UV light induced formation theory of H₂O₂ has also been generally proposed.¹² An extra amount of H₂O₂ will be generated and then participate in the Fenton reaction and as a result it enhances the corresponding removal efficiency. Compared with other typical degradation techniques (e.g., physical absorption, photo-catalytic oxidation, biodegradation, etc.) in waste gas treatment, the UV-Fenton process shows distinct advantages such as in its degradation ability, non-secondary pollution and low operational cost.^13–15 As mentioned above, the UV-Fenton process is considered to be a promising method in landfill gas treatment for the removal of BTEX.

Because VOCs are mineralized through an oxidation reaction with hydroxyl radicals in the aqueous solution, gas pollutants must transfer from the gas phase to the liquid-phase.¹⁶ According to Fick’s (first) law of diffusion, the diffusion flux is proportional to the concentration gradient between the gas phase and the liquid-phase.¹⁷ However, the VOC concentration in landfill gas varies in a comparatively large range and the related concentration adjustment is complex and cumbersome.¹⁸ Thus, enhancement of the effective contact area is more popular. In real applications, an aeration device is usually employed to generate tiny bubbles and then disperse these bubbles in the liquid-phase evenly. With a significant increase in the gas–liquid interfacial area, the corresponding mass transfer efficiency can be apparently enhanced. For instance, the aeration methods have been successfully used in both wastewater treatment and biological reactors to improve the efficiency of the gas–liquid mass transfer process.^19,20 In laboratory research, the bubble column reactor is more likely to be employed. For example, Govindan et al. proposed a biphasic electroreactor comprised of a bubble column for the removal of gaseous benzene and Xia et al. explored the hydrodynamic effects of air sparging on hollow fiber membranes in a bubble column reactor.^21,22 In addition, the Fenton reaction consumes the dissolved VOC and increases the concentration gradient for mass transfer. Thus, a combination of the bubble column and the UV-Fenton reaction is considered feasible in the degradation of VOCs in landfill gas.

This study aimed at presenting an effective method for landfill gas treatment. While a few studies have tried to apply the Fenton process in waste gas degradation, they mainly focused on the removal efficiency of a specific compound and reaction kinetics.^6,23 However, landfill gas has always consisted of various VOCs and they are continuously volatilized in landfills. Thus, we focused on the effective degradation of BTEX which is a group of toxic and typical air pollutants in landfill gas. A bubble column reactor was employed, and its performance on the gas–liquid mass transfer was optimized. The influences of pH, H₂O₂ dosage, Fe²⁺ dosage and dissolved oxygen concentration on the removal efficiency were carefully investigated. To verify the feasibility of this UV-Fenton system in real landfill gas treatment, the degradation of a gaseous mixture of BTEX was performed and its long time continuous operation was also verified.

2. Materials and methods

2.1. Chemicals

All the reagents used in the experiments were of analytical grade. Benzene, toluene, ethyl benzene and o-xylene were obtained from J&K Scientific Ltd (Beijing, China). Hydrogen peroxide (30%, w/v), ferrous sulfate (99%) and other reagents were obtained from Sinopharm Chemical Reagent Beijing Co. Ltd (Beijing, China). Deionized water was used throughout this study.

2.2. Experimental apparatus

As depicted in Fig. 1, the bubble column is a cylindrical reservoir which was constructed out of transparent polymethyl methacrylate (PMMA). A pottery aerator (normally used in a fish tank) was placed in the bottom of the bubble column and a UV lamp (length = 1.0 m, P = 40 W, radiation peak at 254 nm, Creator UV&IR lighting Co., Ltd., China) was inside. The bubble column reactor was wrapped with tinfoil to avoid UV radiation and four observation windows (i.e., S1 to S4 from the bottom to the top in equal intervals) were arranged on the tinfoil to facilitate the observation of gas bubbles.

Fig. 1 The employed UV-Fenton system in the experiments.

A glass bottle (500 mL) containing a liquid reagent was immersed in a water bath, and the volatilization rate of the corresponding reagent was adjusted by controlling the water temperature. With a gas cylinder full of purified air as the carrier gas source, the liquid reagent in the glass bottle was blown by the carrier gas at a constant flow-rate and then mixed with the other flow of pure carrier gas. The desirable inlet gas concentration was determined by the mix ratio of the above mentioned two flow gas which could be controlled with a needle valve and rotameters. During the whole reaction time, the inlet gas was continuously fed to the bubble column with both a constant flow-rate and stable chemical concentration.

2.3. Experimental procedures

The bubble column was filled with 7 L deionized water, and the initial pH value was adjusted by NaOH and HCl as appropriate. After the carrier gas was pumped into the bubble column, H₂O₂ and ferrous sulfate was immediately added (i.e., defined as 0 min of the reaction time). The experiments were carried out at room temperature. No pH adjustment was conducted during the experiments.

The experiments were mainly divided into four steps as follows: (1) in order to investigate the influence of the inlet gas flow-rate on the gas–liquid mass transfer, four different inlet gas flow-rates (i.e., 0.5, 1.0, 2.0 and 3.0 L min⁻¹) were compared. Then, a physical absorption experiment of benzene gas (inlet gas concentration, C_inlet = 420 mg m⁻³) was performed with an optimized inlet gas flow-rate. During the tests, no Fenton reagents were added in the bubble column; (2) the degradation of benzene (C_inlet = 420 mg m⁻³) was performed, and the influence of the initial pH value, H₂O₂ dosage and the molar ratio of Fe²⁺ to H₂O₂ was individually investigated; (3) under the optimized initial conditions of the UV-Fenton process, the role of dissolved oxygen in the solution of the Fenton reagents was investigated by individually employing purified air, oxygen and nitrogen as the carrier gas. In addition, the intermediate products of benzene (C_inlet = 420 mg m⁻³) were also analyzed to confirm their possible degradation pathways; (4) a gaseous mixture of benzene, toluene, ethyl benzene and o-xylene (C_inlet = 420, 683, 430 and 844 mg m⁻³, respectively) was degraded, and the cyclic utilization of this UV-Fenton system was suitably explored.

2.4. Analytical methods

2.4.1. Analysis of the aeration effect. The gas holdup was measured (three times parallel determination) by measuring the level of the aerated liquid (H_AL) and that of the clear liquid (H_CL). Thus, the gas holdup (ε_G) was given as shown in the following equation:²⁴

On the basis of the result achieved above, the specific gas–liquid interfacial area (a) was obtained by the following formulas:

where d₁ (d₂) was the minimum (maximum) axial length of a specific gas bubble; d was the equivalent diameter of each specific gas bubble; N was the number of selected gas bubbles; d_B (Sauter diameter) was the average equivalent diameter of the gas bubbles. Using a digital camera with a 1/3200 second shutter speed, pictures were photographed from observation windows on the bubble column.

2.4.2. Analysis of chemical concentrations. Samples from the gas- and liquid-phases were taken (three parallel samples) at specified time intervals from the sampling ports using syringes. The H₂O₂ concentration was determined by a spectrophotometry method using the potassium titanium(IV) oxalate method (λ = 400 nm).²⁵ The dissolved oxygen was measured by a dissolved oxygen meter (JPSJ-605, INESA Scientific Instrument Co. Ltd). The concentration of BTEX in both the inlet gas and outlet gas was measured by gas chromatography (GC-2014, SHIMADZU, FID detector and DM-5MS capillary column). The carrier gas of gas chromatography was high purity nitrogen at 1 mL min⁻¹ and the temperature program was: 1 min holding at 50 °C, ramping at 10 °C min⁻¹, and final holding for 2 min at 200 °C. Compounds were identified on the basis of their retention times, and the identified compounds were quantified using the external standard calibration procedure. The aqueous solution in the bubble column was sampled and then concentrated by being extracted with acetone. After that, intermediate products were identified by GC/MS (DSQ, Thermal Fisher Scientific).

3. Results and discussion

3.1. Effect of the inlet gas flow-rate

Under a certain test condition (e.g., temperature, pressure, gas concentration, etc.), the efficiency of gas–liquid mass transfer was mainly dominated by both the gas resident time and the gas–liquid interfacial area. The gas holdup (ε_G) was firstly measured and it nearly displayed a linear relation with the increase of the inlet gas flow-rate (Table 1). Thus within this range for the inlet gas flow-rate, the gas resident time of gas bubbles in the bubble column was almost the same. Then, the gas–liquid interfacial area at different inlet gas flow-rates was further measured. The images of the gas bubbles were individually photographed from the four observation windows on the bubble column, and twenty gas bubbles with clear shapes were chosen from the same area of each window (Fig. 2, circled in red). According to eqn (2) and (3), the eighty chosen gas bubbles from the four observation windows were employed to calculate the average equivalent diameter of the gas bubbles (d_B) in the bubble column. As listed in Table 1, the measured d_B values became gradually larger with the increase of the inlet gas flow-rate. Through specific observation of the gas bubble images, the change of the gas bubble size was probably caused by their coalescence. With the increase of the inlet flow-rate, the quantity of gas bubbles in a certain volume of space increased, which would distinctly enhance the probability of coalescence.

Table 1 The measured parameters of gas–liquid mass transfer

Measured parameters	Inlet gas flow-rate (L min^−1)
	0.5	1	2	3
εG ± SD	0.006 ± 0.0003	0.011 ± 0.003	0.022 ± 0.002	0.034 ± 0.003
dB (mm)	1.8	2.3	2.5	3.2

---

Fig. 2 The image of gas bubbles photographed from each observation window (S1 to S4 from the bottom to the top of the bubble column).

With the ε_G and d_B values above measured at different inlet gas flow-rates, the corresponding specific gas–liquid interfacial area (a) was calculated (eqn (4)). If we assume that the coalescence of gas bubbles was avoided and the uniform size of each gas bubble was 1.8 mm (the minimum d_B value in Table 1), the ideal a value at each inlet gas flow-rate could also be calculated. As depicted in Fig. 3a, the ideal a values almost linearly increased from 20 to 113 m² m⁻³. However, the difference between the measured a and ideal a became larger along with an increase of the inlet gas flow-rate. It meant that the degree of bubble coalescence was more serious at a high inlet gas flow-rate. In other words, equal volumes of gases obtained smaller gas–liquid interfacial areas at high inlet gas flow-rates. Although the superficial velocity, which usually represents the treatment ability of a reactor, was proportional to the inlet gas flow-rate, the removal efficiency of waste gas would be negatively influenced when it exceeded a limit range.²⁶ Taking into consideration both the efficiency of the gas–liquid mass transfer and the handling capacity of the UV-Fenton process in this study, an inlet gas flow-rate of 2 L min⁻¹ (a = 52.8 m² m⁻³) was chosen for the experiments. If a higher superficial velocity is desired, more effective methods for gas–liquid mixing should be explored in further research.

Fig. 3 The (a) measured and ideal gas–liquid interfacial areas at different inlet gas flow-rates, and the (b) physical absorption of benzene gas with an inlet gas flow-rate of 2 L min⁻¹ (C_inlet = 420 mg m⁻³).

Before the degradation of BTEX, the physical absorption of benzene gas was performed in advance to confirm the aeration effect of the bubble column reactor. As depicted in Fig. 3b, benzene gas (420 mg m⁻³) was continuously fed into the bubble column. The dissolved benzene concentration increased quickly to 1.0 mg L⁻¹ in the first 50 min of aeration time. At the same time, the outlet gas concentration of benzene also gradually increased. But after 50 min, the increase of the dissolved benzene concentration became slower. Usually, a change in the dissolution rate is attributed to a decrease of the benzene concentration gradient.²⁷ As the aeration time approached 90 min, the outlet gas concentration of benzene became almost the same as the inlet gas concentration. Meanwhile, the dissolved benzene concentration reached 1.2 mg L⁻¹ and remained stable for the rest of the time. With an effective physical absorption, more gaseous contaminants would dissolve in the aqueous solution and be further degraded through the UV-Fenton reaction.

3.2. Influence of initial conditions

3.2.1. Effect of initial pH. As reported by Tokumura et al., both physical absorption and chemical absorption (i.e., Fenton oxidation reaction) existed in the photo-Fenton degradation of toluene gas.²³ Because physical absorption and chemical absorption promoted each other’s proceeding during the whole degradation process, the removal efficiency was normally defined as: (C_inlet − C_outlet)/C_inlet × 100% (C was the gas concentration).

The effect of initial pH on the removal efficiency of benzene gas is illustrated in Fig. 4a. Except for the Fenton reagent with initial pH = 2.1, the removal efficiencies of the other UV-Fenton processes (i.e., pH = 3.0, 4.2, and 5.0) were all maintained above 90% in the first 60 min of the reaction time. Both physical absorption and chemical absorption were considered effective at this period of time. But after 60 min of the reaction time, significant decreases in the removal efficiency were observed for UV-Fenton systems with initial pH = 2.1 and 5.0. They were probably caused by an insufficient amount of ˙OH in the aqueous solution. Besides the consumption of H₂O₂ alongside the reaction processing, extremely acidic conditions usually also cause the reaction to be more inactive and inhibit the formation of ˙OH.²⁸ As a catalytic agent in the decomposition of H₂O₂, Fe²⁺ would be consumed by forming colloidal ferric hydroxide and other ferric hydroxide compounds at high pH values (e.g., pH = 5.0). Also, the scavenger of ˙OH (eqn (5)) was generated at high pH values. In addition, Li and Zhang also found that H₂O₂ was unstable at high pH and would directly decompose to oxygen.²⁹ Thus, a suitable pH value was important for the effective use of H₂O₂. Based on the above mentioned results, the initial pH value of the UV-Fenton process was selected as 3.0.

HO₂⁻ + ˙OH → H₂O + O₂˙⁻ (5)

Fig. 4 Effects of (a) initial pH, (b) H₂O₂ dosage and (c) Fe²⁺ to H₂O₂ molar ratio on the removal efficiencies of benzene (C_inlet = 420 mg m⁻³).

3.2.2. Effect of the initial H₂O₂ dosage. As depicted in Fig. 4b, the effect of the initial H₂O₂ dosage was explored. When the H₂O₂ dosage was 1.1 mmol L⁻¹, the removal efficiency was maintained above 80% in the first 100 min of the reaction time and then decreased considerably. It was mainly caused by both the saturation of physical absorption and the complete use of H₂O₂. As proof of this speculation, no distinct decrease in the removal efficiency was observed for the UV-Fenton systems with an initial H₂O₂ dosage of 5.6 or 11.1 mmol L⁻¹ (Fig. 4b). However, the removal efficiency decreased again when the H₂O₂ dosage further increased. For example, when H₂O₂ dosage increased to 22.2 mmol L⁻¹, the removal efficiencies decreased from 93.5% to 83.6% at 180 min of the reaction time; when it increased to 33.3 mmol L⁻¹, the removal efficiency further decreased to 67.9% in 180 min. As reported by Trovo et al., the decrease in the removal efficiency at very high H₂O₂ dosage was caused by both the self-decomposition of H₂O₂ and its reactions with other compounds (eqn (6) and (7)).³⁰ These reactions would consume H₂O₂ quickly, and further influenced the removal efficiency. As a result, 5.6 mmol L⁻¹ was chosen as the initial H₂O₂ dosage of the UV-Fenton process.

2H₂O₂ → O₂ + 2H₂O (6)

H₂O₂ + ˙OH → HO₂˙ + H₂O (7)

3.2.3. Effect of the Fe(II) to H₂O₂ molar ratio. As illustrated in Fig. 4c, the influence of the Fe²⁺/H₂O₂ molar ratio on benzene gas degradation has been explored. When the molar ratio increased from 0.018 to 0.091, removal efficiencies were distinctly improved during the entire 240 min reaction time. Because ˙OH was mainly generated through the chemical reaction between Fe²⁺ and H₂O₂, an appropriate dosage of Fe²⁺ is helpful for the formation of ˙OH.³¹ But when the molar ratio further increased from 0.091 to 0.138 or 0.184, the removal efficiencies decreased slowly. As reported by Prazeres et al., the decreases were mainly caused by the oxidation reaction between ˙OH and excess Fe(II).³² Thus, a large amount of ˙OH would be used up by Fe²⁺ before the degradation of benzene and its intermediates. Based on the above experimental results, the initial conditions in the following experiments were set as follows: pH = 3.0, C_H2O2 = 5.6 mmol L⁻¹ and the Fe²⁺/H₂O₂ molar ratio = 0.091.

3.3. Effect of dissolved oxygen in the UV-Fenton reaction

As illustrated in Fig. 5a, the removal efficiencies of the UV-Fenton/O₂ and UV-Fenton/Air processes were respectively above 89.4% and 84.9% at 180 min of the reaction time. However, the removal efficiencies of the UV-Fenton/N₂ process were maintained above 90% only in the first 50 min of the reaction time and then quickly decreased to 32% at 180 min of the reaction time. While the change in the H₂O₂ concentration was almost the same in these three UV-Fenton processes, a positive correlation between the removal efficiency and the concentration of dissolved oxygen was observed (Fig. 5). It had been proven that dissolved oxygen would participate in the degradation reactions of the Fenton process. With the involvement of oxygen, there were other ways for the degradation of benzene or its intermediate products.³³ In this study, acetone, acetic acid, hydroquinone, phenol and benzoquinone were also distinctly detected as intermediates in the benzene degradation. This was in accordance with the reported degradation pathway with the involvement of oxygen.³⁴ Thus, a decrease in the removal efficiency of the UV-Fenton/N₂ process should be attributed to the complete use of dissolved oxygen in the aqueous solution (Fig. 5b).

Fig. 5 The (a) benzene removal efficiency, (b) dissolved O₂ concentration and (c) H₂O₂ concentration of the UV-Fenton processes with different kinds of carrier gas. Initial conditions of the UV-Fenton system are: C_{inlet,benzene} = 420 mg m⁻³, pH = 3.0, C_H2O2 = 5.6 mmol L⁻¹ and Fe²⁺/H₂O₂ molar ratio = 0.091.

At the beginning of the reaction, the dissolved oxygen concentration increased from 13.2 mg L⁻¹ to 18.1 mg L⁻¹ (UV-Fenton/O₂), 15.8 mg L⁻¹ (UV-Fenton/Air) and 14.5 mg L⁻¹ (UV-Fenton/N₂), respectively. This increase in dissolved oxygen was probably derived from the decomposition of hydrogen peroxide (Fig. 5c).³⁰ After that, the dissolved oxygen concentration of the UV-Fenton/O₂ and UV-Fenton/Air process decreased and was maintained around 12 mg L⁻¹ and 10 mg L⁻¹, respectively, after 50 min of the reaction time. Based on the removal efficiencies and dissolved oxygen concentrations of the UV-Fenton/O₂ and UV-Fenton/Air process, air was proven capable to supply sufficient oxygen to meet the demands of the UV-Fenton process.

3.4. Degradation of BTEX by the UV-Fenton process and its cyclic utilization

As shown in Fig. 6a, the removal efficiencies of benzene (420 mg m⁻³), toluene (683 mg m⁻³), ethyl benzene (430 mg m⁻³) and o-xylene (844 mg m⁻³) were maintained above 84%, 92%, 96% and 97% in 240 min of the reaction time, respectively. With a delocalized pi (π) bond, benzene usually exhibits relatively better structural stability than many other VOCs.³⁵ In addition, BTEX is a group of typical hydrophobic compounds. However, high water solubility was helpful to enhance the efficiency of the gas–liquid mass transfer and would further improve the removal efficiency of gaseous contaminants. Thus based on the successful degradation of BTEX in this study, it was demonstrated that our proposed UV-Fenton system would be effective in typical VOC degradation.

Fig. 6 The (a) degradation of BTEX and (b) corresponding concentrations of Fenton reagents in the aqueous solution. The inlet gas concentration was 420 mg m⁻³ for benzene, 683 mg m⁻³ for toluene, 430 mg m⁻³ for ethyl benzene and 844 mg m⁻³ for o-xylene. Initial conditions of the UV-Fenton system are: pH = 3.0, C_H2O2 = 5.6 mmol L⁻¹ and Fe²⁺/H₂O₂ molar ratio = 0.091.

As the degradation progresses, most hydrogen peroxide was consumed in the first 150 min of the reaction time (Fig. 6b). Although hydrogen peroxide was the main source of oxidants (i.e., hydroxyl radicals) in the Fenton reaction, the removal efficiency of BTEX was still well maintained after 150 min of the reaction time. Usually, this was attributed to the synergistic effect of UV-light. With the radiation of UV light, part of the organic contaminants would be directly degraded and the utilization level of hydrogen peroxide would also be distinctly improved.³⁶ In addition, hydrogen peroxide would also be continuously generated during the whole degradation process (eqn (8) to (10)).¹² Therefore, the concentration of hydrogen peroxide could be maintained around 0.2 mmol L⁻¹ after 150 min of the reaction time (Fig. 6b).

O₂ + hν → 2O (8)

O + O₂ → O₃ (9)

O₃ + H₂O → O₂ + H₂O₂ (10)

Normally, hydroxyl radicals are generated through the reaction between Fe(II) ions and hydrogen peroxide.³⁷ After that, Fe²⁺ would be oxidized into Fe(III).²³ It could be seen that a large part of Fe²⁺ was oxidized and its concentration decreased rapidly to 10% of its initial concentration in the first 30 minutes of the reaction time (Fig. 6b). But after 30 min of the reaction time, the Fe²⁺ concentration slowly began to rise and it finally reached 0.3 mmol L⁻¹ after 210 min. Several reactions (eqn (11) and (12)) were considered as possibilities for causing the reduction of Fe(III).³⁸

Fe³⁺ + H₂O₂ ↔ Fe(HO₂)²⁺ + H⁺ (11)

Fe³⁺ + HO₂˙ → Fe²⁺ + O₂ + H⁺ (12)

In addition, the reduction reaction between Fe³⁺ and intermediate products was also reported.³⁹ Because the oxidation reaction of Fe²⁺ slowed down along with the decrease of the H₂O₂ concentration, more Fe³⁺ ions were reduced again. Thus, these reactions resulted in the slight increase in the Fe²⁺ concentration after 50 min of the reaction time (Fig. 6b).

Besides a high oxidization ability, an adequate effective lifetime of the UV-Fenton system is also necessary in practical applications. As shown in Fig. 7, a parallel controlled trial of benzene degradation was carried out. Each round of benzene (inlet gas concentration = 420 mg m⁻³) degradation had a 240 min reaction time, the removal efficiencies were measured every 60 minutes and the corresponding average removal efficiency was calculated. Between two continuous rounds, a 30 min aeration of purified air (defined as the interval time) was carried out. In the first experiment (Fig. 6b, no extra H₂O₂), the average removal efficiencies decreased from 95% to 70% after five rounds of reaction. Through the measurement of H₂O₂ concentration, it was found that H₂O₂ was quickly consumed in the period of benzene degradation. However, dissolved benzene would diffuse into the gas-phase at the interval time, and thus physical absorption was almost recovered in the next round. In addition, H₂O₂ was also synthesized with the synergistic effect of UV light at the interval time (eqn (8) to (10)). Therefore, the UV-Fenton process was still effective in the later rounds. In the other experiment (Fig. 7, 10% initial amount added), H₂O₂ was added into the bubble column after the interval time. With the supplement of H₂O₂ at the interval time, the decrease of the average removal efficiency became much slower (decreasing from 95% to 88% after five rounds). Based on the above results, the removal effect of the UV-Fenton process could easily be well maintained in practical applications.

Fig. 7 The cyclic test of the UV-Fenton system (240 min reaction time for each round) with an interval time of 30 min. Initial conditions of the UV-Fenton system are: C_{inlet,benzene} = 420 mg m⁻³, pH = 3.0, C_H2O2 = 5.6 mmol L⁻¹ and Fe²⁺/H₂O₂ molar ratio = 0.091.

4. Conclusion

In the present study, continuous removal of gaseous BTEX by the UV-Fenton process was investigated using a bubble column reactor. After evaluating the aeration effects at different inlet gas flow-rates, the inlet gas flow-rate of the bubble column reactor was optimized at 2 L min⁻¹ and further confirmed by a physical absorption test. Moreover, the most effective initial conditions of the UV-Fenton process (pH = 3.0, C_H2O2 = 5.6 mmol L⁻¹ and Fe²⁺/H₂O₂ molar ratio = 0.091) were chosen after individually exploring their influence on benzene removal efficiencies. In addition, it was confirmed that dissolved oxygen also significantly influenced the removal efficiency by providing extra degradation pathways. Through BTEX degradation, the proposed UV-Fenton system showed distinct characteristics including strong oxidation ability, large handling capacity and cyclic utilization. Taking into consideration pollution features (e.g., complex composition, high concentration and large emission load) in landfills, livestock farms and pharmaceutical factories, the UV-Fenton system is considered to be a promising method for waste gas treatment of air pollution sources.

Acknowledgements

This work was supported by the National Natural Science Foundation of the People’s Republic of China (21277011, 21206009, 21576023) and the Fundamental Research Funds for the Central Universities (FRF-TP-15-080A1).

References

1. Z. Mou, C. Scheutz and P. Kjeldsen, J. Air Waste Manage. Assoc., 2015, 65, 686–698.
2. Z. Zhang, X. M. Wang, Y. L. Zhang, S. J. Lu, Z. H. Huang, X. Y. Huang and Y. S. Wang, Sci. Total Environ., 2015, 511, 792–800.
3. R. Olawoyin, B. Heidrich, S. Oyewole, O. T. Okareh and C. W. McGlothlin, Chemosphere, 2014, 112, 114–119.
4. P. Zhao, J. M. Liu, L. Wang and Q. Yi, Int. J. Environ. Anal. Chem., 2013, 93, 48–60.
5. Y. B. Wang, H. Y. Zhao, M. F. Li, J. Q. Fan and G. H. Zhao, Appl. Catal., B, 2014, 147, 534–545.
6. K. Choi, S. Bae and W. Lee, J. Hazard. Mater., 2014, 280, 31–37.
7. C. M. Zheng, X. Z. Cheng, P. P. Chen, C. W. Yang, S. M. Bao, J. Xia, M. L. Guo and X. H. Sun, RSC Adv., 2015, 5, 40872–40883.
8. N. Xu, Y. Q. Zeng, J. Li, Y. Y. Zhang and W. L. Sun, RSC Adv., 2015, 5, 56832–56840.
9. S. S. Cao, F. F. Kang, P. Li, R. F. Chen, H. Liu and Y. Wei, RSC Adv., 2015, 5, 66231–66238.
10. M. Molkenthin, T. Olmez-Hanci, M. R. Jekel and I. Arslan-Alaton, Water Res., 2013, 47, 5052–5064.
11. N. D. Banic, D. V. Sojic, J. B. Krstic and B. F. Abramovic, Water, Air, Soil Pollut., 2014, 225, 1954.
12. C. Zhao, L. E. Arroyo-Mora, A. P. DeCaprio, V. K. Sharma, D. D. Dionysiou and K. E. O’Shea, Water Res., 2014, 67, 144–153.
13. S. Y. Lu, Y. Ji, A. Buekens, Z. Y. Ma, Y. Q. Jin, X. D. Li and J. H. Yan, Waste Manage. Res., 2013, 31, 169–177.
14. Y. T. Wu, Y. H. Yu, V. H. Nguyen, K. T. Lu, J. C. S. Wu, L. M. Chang and C. W. Kuo, J. Hazard. Mater., 2013, 262, 717–725.
15. W. H. Chen, W. B. Yang, C. S. Yuan, J. C. Yang and Q. L. Zhao, Chemosphere, 2014, 103, 92–98.
16. L. Wang, M. H. Cao, Z. H. Ai and L. Z. Zhang, Environ. Sci. Technol., 2015, 49, 3032–3039.
17. M. Tokumura, M. Shibusawa and Y. Kawase, Chem. Eng. Sci., 2013, 100, 212–224.
18. D. Li, W. J. Lu, Y. J. Liu, H. W. Guo, S. Xu, Z. Y. Ming and H. T. Wang, PLoS One, 2015, 10, e0119305.
19. S. Papoutsakis, F. F. Brites-Nobrega, C. Pulgarin and S. Malato, J. Photochem. Photobiol., A, 2015, 303, 1–7.
20. W. H. Chen, W. B. Yang, C. S. Yuan, J. C. Yang and Q. L. Zhao, Chemosphere, 2014, 103, 92–98.
21. M. Govindan, S. J. Chung, H. H. Moon, J. W. Jang and I. S. Moon, ACS Comb. Sci., 2013, 15, 439–446.
22. L. J. Xia, A. W. K. Law and A. G. Fane, Water Res., 2013, 47, 3762–3772.
23. M. Tokumura, R. Nakajima, H. T. Znad and Y. Kawase, Chemosphere, 2008, 73, 768–775.
24. R. B. Moruzzi and M. A. P. Reali, Water Sci. Technol., 2010, 61, 253–262.
25. M. Tokumura, A. Ohta, H. T. Znad and Y. Kawase, Water Res., 2006, 40, 3775–3784.
26. T. A. Alsmari, J. R. Grace and X. T. T. Bi, Chem. Eng. Sci., 2015, 123, 49–56.
27. L. Zhang, W. Wang, X. J. Ju, R. Xie, Z. Liu and L. Y. Chu, RSC Adv., 2015, 5, 5638–5646.
28. J. M. Monteagudo, A. Duran, M. Aguirre and I. S. Martin, J. Hazard. Mater., 2011, 185, 131–139.
29. Y. M. Li and A. Zhang, Chemosphere, 2014, 105, 24–30.
30. A. G. Trovo, T. F. S. Silva, O. Comes, A. E. H. Machado, W. B. Neto, P. S. Muller and D. Daniel, Chemosphere, 2013, 90, 170–175.
31. B. D. Wu, S. J. Zhang, X. C. Li, X. T. Liu and B. C. Pan, Chemosphere, 2015, 131, 55–62.
32. A. R. Prazeres, F. Carvalho and J. Rivas, J. Environ. Manage., 2013, 129, 199–205.
33. Y. X. Du, M. H. Zhou and L. C. Lei, J. Hazard. Mater., 2007, 139, 108–115.
34. B. Utset, J. Garcia, J. Casado, X. Domenech and J. Peral, Chemosphere, 2000, 41, 1187–1192.
35. H. B. An, S. H. Park, S. H. Jhurng, J. S. Jurng, G. N. Bae, J. K. Jeon and Y. K. Park, J. Nanosci. Nanotechnol., 2015, 15, 454–458.
36. T. M. Chen, Y. Q. Zhang, J. L. Yan, C. Ding, C. T. Yin and H. Liu, RSC Adv., 2015, 5, 12638–12643.
37. L. Yu, X. F. Yang, Y. S. Ye and D. S. Wang, RSC Adv., 2015, 5, 46059–46066.
38. C. C. Jiang, Z. Gao, H. L. Qu, J. W. Li, X. X. Wang, P. Li and H. Liu, J. Hazard. Mater., 2013, 250, 76–81.
39. M. Gencten and A. Ozcan, Chemosphere, 2015, 136, 167–173.

---