Yanghaichao LIU (劉楊海超),Liping LIAN (連莉萍),Weixuan ZHAO(趙瑋璇),Renxi ZHANG (張仁熙) and Huiqi HOU (侯惠奇)

1 Shanghai Key Laboratory of Atmospheric Particle Pollution and Prevention (LAP3),Institute of Environmental Science,Fudan University,Shanghai 200433,People’s Republic of China

2 National Engineering Laboratory for VOCs Pollution Control Material &Technology,University of Chinese Academy of Sciences,Beijing 100049,People’s Republic of China

Abstract

Keywords:plasma catalytsis system,chlorinated VOCs,MnOx/Al2O3 catalysts

1.Introduction

Chlorinated volatile organic compounds (VOCs) are a common contaminant widely detected in environmental media.Compared with other VOCs,chlorinated VOCs are more toxic and tend to accumulate in the environment [1].Therefore,chlorinated VOCs have been included on hazardous chemical lists in many countries [2].In recent decades,the study of chlorinated VOCs degradation has mainly focused on the adsorption method [3–8],photocatalysis [9],thermal treatment[10,11],and catalytic oxidation[12–17].However,the adsorption method is a nondestructive method,while photocatalysis has a slow reaction rate and thermal treatment requires high temperatures as well as huge energy consumption.The catalytic oxidation method is also limited by a high and narrow temperature window for the use of catalysts.Moreover,some studies indicated that the catalytic oxidation of chlorobenzene at high temperatures can produce polychlorinated benzene [18,19].

Many studies have shown that nonthermal plasma technology is an efficient VOCs treatment technology that has a good effect on many kinds of VOCs [20].However,nonthermal plasma still has many disadvantages,such as low energy efficiency,low selectivity of COx,and undesirable byproducts [21–24].Therefore,the hybrid technologies of nonthermal plasma technology coupled with other technologies have been studied[25–27].Among these hybrid systems,plasma catalysis technology can overcome these limitations and improve the removal rate of contaminants and mineralization rate at the same time because the gas-phase reaction and catalyst surface reaction induce a synergistic effect of plasma and catalysts [28–32].Therefore,plasma catalysis technology and the degradation of chlorinated VOCs has received increased attention [33].Many kinds of catalysts have been used in the plasma catalytic system.Among the various catalysts,Mn-based catalysts are known for exhibiting high activity in catalytic combustion of chlorinated VOCs[34–38].In addition,some studies have demonstrated that MnOxcatalysts show good resistance to chlorine poisoning[39].Furthermore,MnOxcatalysts exhibit high activity in the ozonation of chlorinated VOCs [40].These advantages of Mn-based catalysts make them a common catalyst for plasma catalytic degradation of chlorinated VOCs [41,42].

Figure 1.Schematic diagram of the experimental setup.

However,current studies on chlorinated VOCs using a plasma-coupled catalytic system mainly focus on chlorinated aliphatic hydrocarbons,such as trichloroethylene and chloroform.Few studies have been carried out on chlorinated aromatic hydrocarbons using a plasma-coupled reactor especially with MnOxcatalysts [43].Moreover,some byproducts produced in dielectric barrier discharge (DBD) coupled with a catalysis system are more toxic than the original chlorinated pollutants [44],so it is important to clarify the formation pathways of reaction products in this technology.Furthermore,the mechanism of degradation of chlorinated aromatic hydrocarbons and the effect on the byproducts in the coupled plasma catalysis system are not clear.

In this paper,chlorobenzene,the most typical chlorinated aromatic hydrocarbon,was selected as the target contaminant.The effect of chlorobenzene degradation using a DBD coupled with an MnOx/Al2O3catalysts system was studied.The removal rate of chlorobenzene,the selectivity of CO and CO2,and the concentration of O3were investigated.Furthermore,the byproducts and the possible reaction mechanism are also discussed in this paper.

2.Experiment

2.1.Catalyst preparation and characterization

The MnOx/γ-Al2O3catalysts were prepared by the impregnation method.Manganese acetate solution (Mn(CH3COO)2·5H2O)was chosen as the precursor and γ-Al2O3powder was added to a certain concentration of excessive manganese acetate solution.After 24 h of impregnation,the excess solution was filtered out and the remaining slurry was dried at 80°C for 12 h.The samples were then calcined at 400°C for 4 h.After grinding and sieving,the prepared MnOx/γ-Al2O3catalysts with particle sizes ranging from 0.25 mm to 0.425 mm(40–60 mesh)were selected for the experiment.In addition,for the control group,distilled water was used instead of the manganese acetate solution to prepare unloaded Al2O3particles via the same process.

Inductively coupled plasma optical emission spectroscopy (ICP-OES) measurements were carried out to determine the load of Mn on the catalyst using an inductively coupled plasma optical emission spectrometer (Perkin Elmer Optima 5300DV).N2adsorption and desorption experiments were performed on a Micromeritics TriStarII 3020 instrument to measure the Brunner?Emmet?Teller (BET) surface areas of the catalysts.The x-ray diffraction (XRD) patterns of the catalysts were collected on an x-ray diffractometer (Bruker D8 Advance) using Cu Kα radiation (λ=0.1542 nm) at 40 kV and 40 mA in a 2θ range of 10°C–90°C.X-ray photoelectron spectroscopic (XPS) was performed on a Thermo ESCALAB 250XI.

2.2.Experimental setup

A schematic diagram of the experimental setup is shown in figure 1.The experiment was carried out at room temperature and atmospheric pressure.Air (79% N2+21% O2) was divided into two streams:one stream was mixed with another stream of air in a buffer tank after passing through a chlorobenzene bubbling bottle that was placed in a water bath(25°C±1°C).The flow rate was controlled by a mass flow controller (MFC;Beijing Sevenstar Flow Co.,Ltd).Theconcentration of chlorobenzene was 200 ppm (approximately 1000 mg m–3) and the total gas flow rate was 3 l min–1.The DBD reactor that was used in the experiment was a coaxial cylinder reactor with a discharge gap of 3 mm.The inner diameter of the outer quartz glass cylinder was 26 mm,the wall thickness was 1.5 mm,and 40 mm long copper foil was attached to the outer wall as the outer electrode.The inner cylinder was a quartz glass cylinder with an outer diameter of 20 mm and was filled with graphite powder as the inner electrode.Catalysts of 0.5 g MnOx/Al2O3(or Al2O3powder) were packed in the discharge area.The mixed chlorobenzene gas entered the upper inlet of the reactor and flowed out from the bottom of the reactor.To exclude the effect of catalyst adsorption on the removal of chlorobenzene,the experiment was carried out after the catalyst bed reached adsorption equilibrium (the concentration of chlorobenzene at the inlet was equal to that at the outlet).

Table 1.The relationship between applied voltage and the SED of the reactor.

The reactor was powered by an AC power supply that provided a sinusoidal alternating voltage varying from 3.6 kV to 6 kV at a frequency of 10–20 kHz.The voltage and power applied were measured via the voltage-charge Lissajous figure with a 200 MHz digital phosphor oscilloscope (Tektronix,TDS2024B,USA)connected to a 1000:1 high-voltage probe (Tektronix P6015A,USA).The current was obtained by measuring the voltage of a resistor,and the Lissajous figure was measured at the discharge electrode and a 0.47 μF equivalent capacitor.

The specific energy density (SED;J l–1) of plasma is defined as the energy obtained per unit of gas:

where P(W)is the power consumed by the DBD reactor and Q(l min–1)is the gas-flow rate.The relationship between applied voltage and the SED of the reactor is shown in table 1.

In the experiment,the concentration of chlorobenzene was determined by a gas chromatograph (GC9790,FULI Instruments,Zhejiang China)with a flame ionization detector(FID).The concentration of CO and CO2were determined by a gas chromatograph (GC930,Haixin Instruments,Shanghai China) with a thermal conductivity detector (TCD).The concentration of O3was determined by the iodometric method.The removal rate of chlorobenzene and the selectivity of CO,CO2,and COxwere calculated from the following formulas:

Figure 2.Adsorption curves of the MnOx/Al2O3 catalysts and Al2O3 powder.

where[CB]in(ppm)and[CB]out(ppm)are the concentrations of chlorobenzene at the inlet and outlet,respectively,and[CO](ppm) and [CO2](ppm) are the volumetric concentrations of CO and CO2,respectively,at the outlet.

3.Results and discussion

3.1.Characterization of catalysts

The amount of manganese in the MnOx/Al2O3catalyst was determined by ICP-OES.According to the results,the amount of manganese was 4.5 wt%.The structural properties of the MnOx/Al2O3catalyst and Al2O3powder were determined by N2adsorption/desorption experiments.The adsorption curves in figure 2 show that the adsorption curves of both materials belong to Type II isotherms.In addition,hysteresis loops appear when the relative pressure is between 0.7 and 1.0,which indicates that capillary condensation occurs in both materials,and the results show that the two samples are typical porous adsorbents.Table 2 shows the specific surface area,pore volume,and average pore size of these two materials.The specific surface area of the Al2O3powder is 168.6 m2g?1and the pore volume is 0.95 cm3g?1.After Mn loading,the surface area and pore volume of the MnOx/Al2O3catalysts were reduced to 146.3 m2g?1and 0.77 cm3g?1,respectively,because the introduced Mn covered part of the pores.

Table 2.Physicochemical properties of the of MnOx/Al2O3 catalyst and Al2O3 powder.

Figure 3.XRD patterns of the MnOx/Al2O3 catalysts and the Al2O3 powder.

Figure 4.XPS spectra of Mn 2p in the MnOx/Al2O3 catalysts.

Figure 3 shows the XRD patterns of the MnOx/Al2O3catalyst and the Al2O3powder.Compared with the diffraction pattern of Al2O3,the diffraction signals of the MnOx/Al2O3catalyst at 2θ=28.8°,36.0°,59.8° and 64.6° can be attributed to the formation of Mn3O4crystals (JCPDS 24-0734).Figure 4 shows the XPS spectrum of Mn 2p in the MnOx/Al2O3catalyst;the 2p spectra of Mn clearly show the spin orbital splitting.The peaks of 642.3 eV and 654.1 eV correspond to the Mn 2p1/2peak and Mn 2p3/2peak,respectively,and the energy difference between these two peaks ΔE (2p3/2–2p1/2) was 11.6 eV,which can be assigned to the mixed-valence manganese oxide Mn3O4[45].Additionally,the valence distribution of Mn can be identified from the peak splitting of the Mn 2p peak.The Mn 2p3/2peak was deconvoluted into two peaks of 654.5 eV and 653.2 eV;similarly,the Mn 2p1/2peak was deconvoluted into two peaks of 642.9 eV and 641.6 eV[46,47].An additional peak of 646.1 eV corresponds to the satellite peak of Mn 2p3/2.This result indicates that Mn is present in two oxidation states of Mn3+and Mn2+in MnOx/Al2O3catalysts,which is consistent with the presence of Mn3O4and corroborates the XRD results.Furthermore,according to the deconvolution results,the Mn2+/Mn3+ratio in the catalysts is calculated to be 2.11:1,which is very close to the predicted spinel ratio value of 2 [48].

3.2.Temperature change in the DBD reactor

After 15 min of discharge at the SED of 1350 J l–1,the temperature in the discharge area was approximately 105°C.In order to explore the effect of temperature on the reaction,an experiment was carried out at 105°C to study the degradation of chlorobenzene on the catalysts without discharge,and the results showed that there was no obvious degradation of chlorobenzene by the catalysts without discharge.Therefore,in this experiment,the direct thermal effect of gas temperature change on the degradation of chlorobenzene can be ignored.

3.3.Removal rate of chlorobenzene

Figure 5 shows the chlorobenzene removal rate as a function of the SED in the single DBD reactor and the combined plasma catalysis reactors.As shown in figure 5,compared with the single DBD reactor and the reactor packed with Al2O3powder,the removal rate of chlorobenzene in the reactor packed with MnOx/Al2O3catalysts was the highest whether at high or low SED.Especially,in the case of low SED,the removal rate in the reactor packed with MnOx/Al2O3catalysts was obviously better than that of the other two reactors.When the SED was 750 J l–1and the concentration of chlorobenzene was 200 ppm,the removal rate in the reactor packed with MnOx/Al2O3catalysts was 70.6%,while that of the other two reactors was only 53.8%and 38.9%,respectively.However,with the increase of SED,the differences in the removal rates between the reactor packed with MnOx/Al2O3catalysts and the other two reactors became very small,and even the removal rates in the reactor packed with Al2O3powder was almost the same as that in the single DBD reactor.

Figure 5.Effect of SED on chlorobenzene removal rate in the single DBD reactor and the combined plasma catalysis reactor.

These conclusions indicate that,at low SED,the adsorption effect of the Al2O3powder and MnOx/Al2O3catalysts increased the residence time and the concentration of chlorobenzene on these surfaces.This behavior greatly increases the reaction probability between chlorobenzene and active particles generated by discharge.In addition,MnOx/Al2O3catalysts can also directly oxidize chlorobenzene.According to the Mars–van Krevelen mechanism,chlorobenzene adsorbed on the MnOxsurface was oxidized by oxygen atoms in the catalysts and the oxygen vacancies can be replenished by capturing oxygen molecules or catching oxygen from the decomposition of ozone [49–52].However,with increasing voltage and SED,the electric intensity in the reaction area increases,which can increase the number of energetic electrons and the average electronic energy,and further increases the number of active particles and radicals.On this occasion,the main contribution to the decomposition of chlorobenzene was the direct destruction by plasma,while the catalysts mainly contributed to the oxidation of intermediates.Therefore,when the SED was higher than 1100 J l–1,the difference between the removal rate of the reactor packed with catalysts and that of the other two reactors did not change significantly.

3.4.Ozone concentration

Ozone is considered to be a harmful byproduct in DBD technology.Figure 6 presents a comparison of O3concentration between the single DBD reactor and the two combined plasma catalysis reactors.As shown in figure 6,there was no obvious difference in the three reactors at low SED.With an increase of SED,the enhancement of discharge produced more ozone and the difference between all three reactors became more and more obvious.However,the lowest ozone concentration was detected in the reactor packed with MnOx/Al2O3catalysts and the highest ozone concentration was found in the reactor packed with Al2O3powder.This indicated that,in the reactor packed with Al2O3powder,the adsorption effect of Al2O3could increase the existence time of oxygen atoms,resulting in the production of more ozone.However,although the MnOx/Al2O3catalysts also had the adsorption effect,the active component MnOxcould degrade ozone efficiently and led to a significant reduction in ozone concentration as a result of the lowest ozone concentration.These experiments show that ozone generated by discharge can be decomposed on the surface of MnOx/Al2O3catalysts to produce active oxygen atoms and oxygen;similar results have also been reported [53].The reaction formula is as follows:

Figure 6.Effect of SED on the concentration of O3 in the single DBD reactor and the combined plasma catalysis reactor.

where*represents the active sites on the catalysts.

In DBD,oxygen atoms produced by discharge can easily combine with oxygen molecules to produce ozone.Meanwhile,the oxidation of chlorobenzene also needs to consume the active oxygen atoms.On the one hand,when the MnOx/Al2O3catalysts were packed in the DBD reactor,the catalysts directly used ozone to oxide chlorobenzene;on the other hand,the ozone produced by discharge was decomposed back into oxygen atoms and then participated in the degradation of chlorobenzene.Therefore,the packing of the MnOx/Al2O3catalysts cannot only reduce the ozone concentration produced in the DBD reactor,but also improve the energy utilization efficiency of the reactor.

3.5.Selectivity of CO and CO2

In the degradation of chlorobenzene,the selectivity of COxis an important factor,which can reflect the degree of oxidation of chlorobenzene in the reactor.The selectivity of COx,CO,and CO2as a function of SED in the single DBD reactor and two combined plasma catalysis reactors is shown in figures 7(a)–(c).Comparing the differences between these three reactors,the highest selectivity of COxwas found in the reactor packed with MnOx/Al2O3catalysts with the highest selectivity of CO2and the lowest selectivity of CO.Meanwhile,the difference between this reactor and the other two reactors in the selectivity of CO2is greater than that in the selectivity of CO.

In addition,there was no significant difference in both selectivity of CO and CO2in the other two reactors.These results further indicate that only packing Al2O3powder had no effect on the selectivity of COx.However,the introduction of MnOxnot only oxidized more chlorobenzene or intermediates to produce COx,but also converted more CO to CO2.In the single DBD reactor and the reactor packed with Al2O3,the active oxygen consumed by the oxidation of chlorobenzene was generated by discharge,while in the reactor packed with catalysts,MnOx/Al2O3catalysts can provide two additional oxidation pathways for intermediates except for the oxidation by active oxygen produced by discharge.The first is oxidation by oxygen atoms produced by ozone decomposition and the second is direct oxidation on the surface of the catalysts.These two pathways allow intermediates adsorbed on the catalyst surface to be further oxidized to form more CO2.Therefore,DBD coupled with MnOx/Al2O3catalysts can obviously improve the mineralization of chlorobenzene.

Figure 7.Effect of SED on the selectivity of COx(a),CO(b),and CO2(c)in the single DBD reactor and the combined plasma catalysis reactor.

3.6.Byproducts and reaction mechanism

To obtain a clear understanding of the degradation process of chlorobenzene,the final products in the single DBD reactor and the coupled MnOx/Al2O3catalyst reactor were analyzed by a gas chromatography-mass spectrometer (GC-MS).The results shown in table 3 indicate that the products included some ring-opening products,some aromatic hydrocarbons,and dichlorodiphenyl ethers.It is important to note that the position of substituents is not certain for the aromatic compounds.

The degradation of chlorobenzene by DBD usually goes through three processes:electron impact dissociation,gasphase radical reactions,and ion-molecule reactions.Previous studies reported that electron impact was the most important channel during the discharging periods and was responsible for the initial reaction for VOCs decomposition [54].In the chlorobenzene molecule,the bond energy of the C-Cl bond is approximately 4.2 eV [55].In plasma discharge,the C-Cl bond breaks first to form a Cl atom and phenyl.In the benzene ring,the bond energies of the ortho,meta,and para C-H bonds are 3.5 eV,4.1 eV,and 4.8 eV[56],respectively.These C-H bonds can also be destroyed by high-energy electrons.Meanwhile,N2molecules in the carrier gas can be converted to NO under discharge and further form NO2.In the discharge area,some active particles such as Cl atoms,OH radicals,and NO2(nitroxyl radical) can react with phenyl to form these aromatic compounds.All these intermediates can be attacked by high-energy electrons or oxidized by reactive species.The reaction formulas are as follows:

where X represents arbitrary substituents,such as OH,Cl,and NO2.

In addition,two substances appeared in the byproduct analysis of the two reactors,oxalic acid(HOOC–COOH)and butene diacid (HOOC–C=C–COOH).The peak signals of these two substances are stronger than the other signals,so we surmise that the opening of the benzene ring mainly forms two carbon chains with four and two carbon atoms.Similar decomposition mechanisms have also been reported[57].The reaction formula is as follows:

Table 3.Byproducts analysis results.

In these small molecules,the C=C bond can be oxidized continuously and eventually completely oxidized to CO,CO2,and H2O.Another noteworthy point is that,during the decomposition of the oxalic acid molecule,CO and CO2at a ratio of 1:1 can be formed.The reaction formulas are as follows:

Furthermore,in the previous discussion,the ratio of CO and CO2produced in a single DBD reactor was between 1.0 and 1.1,which indicates that most of the degradation pathways of chlorobenzene in DBD are carried out through the decomposition of oxalic acid.

Comparing the differences of byproducts between these two reactors,the reactor packed with catalysts is more easily able to produce polychlorinated compounds.Catalysts provided a platform for Cl atoms accumulation due to adsorption.When the intermediates are adsorbed on the catalysts,the Cl atoms enriched on the catalyst surface can easily replace the H atoms to form polychlorinated substitutes.Therefore,although the DBD-coupled MnOx/Al2O3catalysts reactor can efficiently removal chlorobenzene,the effect of the catalysts on the production of harmful byproducts cannot be ignored.

4.Conclusions

Compared with the single DBD reactor and the reactor coupled with Al2O3powder,the reactor coupled with the MnOx/Al2O3catalysts has the highest removal rate,the highest selectivity of COx,and the lowest ozone concentration.The packing of MnOx/Al2O3catalysts can obviously oxidize chlorobenzene and intermediates to produce more CO2,especially in the case of low SED.Moreover,the catalysts also efficiently decomposed ozone and provided more oxygen atoms for the oxidation of reactants.With the increase of SED,the degradation of chlorobenzene was mainly destroyed by the direct discharge.In addition,based on the results of byproducts analysis,in the decomposition of chlorobenzene the opening of benzene rings forms unsaturated hydrocarbons with four and two carbon atoms.Despite many advantages of DBD coupled with MnOx/Al2O3catalysts,we also found that the packing of catalysts increased the production of polychlorinated compounds.

Acknowledgments

The financial support for this research was provided by National Natural Science Foundation of China (No.21577023),the Special Research Project on Causes and Control Technology of Air Pollution (Nos.2017YFC0212905),and the Science and Technology Innovation Action Project Supported by the Science and Technology Commission of Shanghai Municipality(No.18DZ1202605).The authors thank Jianyuan Hou and Tianye Li for their help in the research work.

ORCID iDs