Jingying Xu *,Yue Lyu ,Jiankun Zhuo ,Yishu Xu ,Zijian Zhou ,Qiang Yao,*

1 Department of New Energy Science and Engineering,School of Energy and Power Engineering,Huazhong University of Science and Technology,Wuhan 430074,China

2 State Key Laboratory of Coal Combustion,Huazhong University of Science and Technology,Wuhan 430074,China

3 Key Laboratory for Thermal Science and Power Engineering of Ministry of Education,Department of Energy and Power Engineering,Tsinghua University,Beijing 100084,China

4 Department of Power Engineering and Mechanics,School of Energy and Power Engineering,Huazhong University of Science and Technology,Wuhan 430074,China

Keywords:Volatile organic compounds Coal combustion Ozone formation potential Coal-fired power plant On-site measurement

ABSTRACT On-site measurements of volatile organic compounds(VOCs)in different streams of flue gas were carried out on a real coal-fired power plant using sampling bags and SUMMA canisters to collect gas samples,filters to collect particle samples.Gas chromatography-flame ionization detector/mass spectrometry and gas chromatography-mass spectrometry was the offline analysis method.We found that the total mass concentration of the tested 102 VOC species at the outlet of wet flue gas desulfuration device was (13456 ± 47) μg·m-3,which contained aliphatic hydrocarbons (57.9%),aromatic hydrocarbons(26.8%),halogen-containing species (14.5%),and a small amount of oxygen-containing and nitrogencontaining species.The most abundant species were 1-hexene, n-hexane and 2-methylpentane.The top ten species in terms of mass fraction(with a total mass fraction of 75.3%)were mainly hydrocarbons with a carbon number of 6 or higher and halogenated hydrocarbons with a lower carbon number.The mass concentration of VOC species in the particle phase was significantly lower than that in the gas phase.The change of VOC mass concentrations along the air pollution control devices indicates that conventional pollutant control equipment had a limited effect on VOC reduction.Ozone formation potential calculations showed that aromatic hydrocarbons contributed the highest ozone formation(46.4%)due to their relatively high mass concentrations and MIR (maximum increment reactivity) values.

1.Introduction

Coal still plays a very important role in energy applications in lots of countries,especially in China [1].The combustion process of coal can produce particulate matter (PM),sulfur dioxide (SO2),nitrogen oxide (NOx),mercury (Hg) and organic compounds(OCs).At present,relevant studies about the emission characteristics of PM [2–4],SO2[5],NOx[6] and Hg [7] in the coal-fired flue gas have been carried out continuously for many years,and relative mature pollutant control technologies have been formed.However,the emission characteristics of OCs in coal combustion process has not received enough attention.

OCs,including a large number of individual chemical species,exist throughout the atmosphere in gas phase,particle phase or between the two phases.They are divided into volatile organic compounds (VOCs),intermediate-volatility organic compounds(IVOCs),semi-volatile organic compounds (SVOCs),etc.by vapor pressure [8].Among them,VOCs exist only in the gas phase and can react with NOx,ozone and OH radical in the atmosphere under the condition of sunlight [9],leading to the depletion of stratosphere ozone [10],the generation of ground-level ozone [11–14]and secondary organic aerosols [15–18],and the enhancement of photochemical smog and global greenhouse effect [19].Some VOCs,such as aldehydes and low-ring-number polycyclic aromatic hydrocarbons,are shown to be carcinogenic and toxic.They may also have compounding short-term and long-term health effects at low concentrations [20].

Studies have indicated that the use of fossil fuel is one of the primary sources of anthropogenic emissions of non-methane volatile organic compounds (NMVOCs) in China [21].In winter,the proportion of stationary combustion sources will increase further[22].However,most of the existing research only considered VOC emissions from vehicle exhaust,biomass burning,solvent/-paint use and so on.VOCs emitted from coal combustion are less discussed.Information on the specific categories and concentrations of VOCs emitted directly from coal-burning processes,rather than those in the atmosphere surrounding coal-fired equipment is particularly limited.

Previous studies have discussed the emission characteristics of VOCs from coal-fired process by sampling directly in the stacks.Garciaet al.[23] sampled 25 VOCs with Tenax-TA and Carbotrap tubes from 5 coal-fired power plants(130–510 MW)and analyzed them using thermal desorption and gas chromatography-mass spectrometry (GC/MS).It was found that the total concentration of the sampled 25 species was 360.5 μg·m-3.And aldehydes,aliphatic,aromatic and halogenated hydrocarbons were the main components.Fernández-Martínezet al.[24,25] used a volatile organic sampling train (with carbotrap B sorbent) and GC/MS to study the emissions of 16 VOCs from 5 Spanish coal-fired power plants (80–550 MW).The total concentration of the 16 VOCs was between 88.6–447.4 μg·m-3and BTEX (and other alkylbenzenes)and heptane were the predominant species.dos Santoset al.[26]sampled 36 VOCs from a coal-fired power plant (Jorge Lacerda)in Latin America with an installed capacity of 832 MW using XAD-2 and activated charcoal tubes.The total concentration of these VOC species was 498.8 μg·m-3and benzene and hexadecane were released in high concentration.Notably,they observed that high contents of benzo[a]pyrene and pyrene were related to the PM in the atmosphere around the power plant.Pudasaineeet al.[27] also used a volatile organic sampling train and studied the emission characteristics of 8 VOCs from coal and oil-fired power plants.It showed benzene and halogenated hydrocarbons were the main components.And ESP and FGD process had no effect on the reduction of VOC emissions.Shiet al.[28]measured 107 VOCs from four coal-fired process including coke production,iron smelt,thermal power plant and heating station plant in the northeast of China using SUMMA canister and GC/MS.For thermal power plant,the total concentration of the measured 107 species was 16287 μg·m-3and oxygen-containing compounds (ketones,alcohols,acetates,etc) were considered as the representative species.For non-methane hydrocarbons (NMHCs),C6–C7 contributed to 64.4% of the total NMHCs,with toluene andn-hexane holding the highest concentrations.Yanet al.[29]measured 57 VOCs from a coal-fired power plant in North China using sampling bags and GC/MS.The total emission factor of the tested species is 0.023 g·(kg coal)–1(or 0.88 g·GJ-1),and 1-butene,styrene,n-hexane and ethylene were the most abundant species.

Previous studies have indicated that sampling methods have a great effect on the mass concentrations of VOCs emitted from coal-fired process.In addition,in recent years,the coal-fired power plants in China have generally adopted low-nitrogen combustion technology and are equipped with relatively sophisticated air pollution control devices (APCDs) to synergistically remove the conventional air pollutants (SO2,NOxand PM).Although it has been suggested that APCDs may have a positive effect on VOC reduction[30],most of the above-mentioned on-site measurements were located in the stacks,the impact of different air pollution control devices on VOC emissions still needs further investigation.

Fig.1.Diagram of the sampling site distribution.

In this study,we qualified and quantified VOC emissions from coal-fired boilers with low-nitrogen combustion technology in a real power plant,and investigated the effects of the existing APCDs,including selective catalytic reduction (SCR),fabric filter (FF),and wet flue gas desulfuration (WFGD),on VOC reduction.The chemical reactivity of individual VOCs was also discussed to evaluate its environmental impact.

2.Experimental

2.1.The tested power station unit

In this study,gas samples and particle samples were collected from a boiler in a power plant situated in the middle of China.The installed capacity of the power generation unit was 220 MW.The distribution of the sampling sites was exhibited in Fig.1,and Table 1 lists the detailed parameters of the flue gas at each sampling sites.The coal used during the sampling period was a bituminous coal,and Table 2 lists the results of its proximate and ultimate analysis.

Table 1Temperature,O2,dust,SO2 and NOx content of the flue gas at each sampling sites

Table 2Proximate analysis and ultimate analysis of the coal

2.2.Sampling of VOCs

The gas samples were collected using both sampling bags(with a volume of 3 L)and SUMMA canisters(with a volume of 3 L).The sampling bags were cleaned through high purity nitrogen before using.All parts of the sampling device were made of Teflon.During sampling,the sampling probe was inserted into the sampling hole.The end of the sampling probe is as close as possible to the center of the flue and faces the flue gas.The transfer line was heated to maintain the surface temperature to avoid particle condensation.A vacuum box and a sampling pump were used to collect particle-free flue gas into sampling bags.Prior to sampling,the entire sampling system including the sampling bag was flushed with flue gas and then evacuated at least 3 times to reduce sample loss due to adsorption on the inner surface of the sampling system.After continuous sampling at a certain sampling site,the entire sampling system will be flushed with high purity nitrogen.All samples collected by sampling bags were stored in a dry and cool place and analyzed within 3 days.

At the same time,we also used SUMMA canisters to collect gas samples for comparison with sampling bags.Because of the low concentration limit of the gas sample required by SUMMA canisters,we employed a two-stage dilution sampling system.The dilution ratio of the first stage was 40 to 60 or 80 to 120 and that of the second stage ranged from 2 to 2000.In this study,we used a typical dilution ratio of approximately 200.The sampling process of each sampling site was also carried out in three steps,i.e.pre-rinsing,sampling and post-rinsing.All samples collected by SUMMA canisters were analyzed within 7 days.

To investigate the distribution of VOCs in the gas and particle phase,solid samples were sampled simultaneously when sampling gas samples by using a smoke sampler.Due to the difference in particle concentration,glass fiber filter cartridge and filter membrane were used for particle sampling before and after the precipitator.For each sampling site,2–3 parallel samples were collected.The filter cartridge or filter membrane was weighed before and after each experiment.And then,a certain amount of solid samples from the filter cartridge or a filter punch with a certain amount of solid sample was placed in an empty sorbent tube.Both ends were plugged with quartz wool.Before use,both the empty sorbenttubes and the quartz wool should be conditioned at 325°C through ultra-pure N2at 0.5 L·min-1for 3 h.All tubes were analyzed within 7 days.

2.3.Sample analysis and QC/QA

Firstly,the gas samples were concentrated in a preconcentrator to concentrate VOCs and remove H2O and CO2.Then,thermal desorption was performed,and the desorbed VOCs was carried by ultra-pure He to a gas chromatography-flame ionization detector/mass spectrometry (GC-FID/MS,TH-300B,Agilent,USA).In this analysis system,VOCs were separated using a DM-PLOT MS 5A column (15 m × 0.32 mm × 30 μm) and a DB-624 column(60 m × 0.25 μm × 1.4 μm),wherein the lower molecular weight components (C2–C5) were analyzed by FID whereas the higher molecular weight components including oxygenated and halogenated compounds were analyzed by MS.The initial temperature was held at 35 ℃for 3 min.Then,it was risen to 180°C at a rate of 6 °C·min-1,and kept for 5 min until all target species were desorbed.The MS was used in selected ion monitoring (SIM) mode,scanning in a mass range (m/z) from 35 to 200 and the ionization method was electron impact(EI,70 eV)with a source temperature of 230 °C.Target species were qualified according to the retention time and mass spectrum,and quantified through multipoint external calibration method.Table S1 displays the measured 102 species in this paper.Standard gas mixtures were PAM calibrating gas(1.0 cm3·m-3,including 57 compounds,Spectra Gases,USA) and VOC calibrating gas (1.0 cm3·m-3,including 45 compounds,Spectra Gases,USA),which were diluted dynamically to 0.6,1.0,2.4,4.0 and 8.0 mm3·m-3to prepare calibration standards.

To measure the VOCs in solid samples,the sorbent tubes containing solid samples were injected by TurboMatrix 100 thermal desorber (Perkin-Elmer,USA).The tubes were heated to 325 °C for 3 min.VOCs were desorbed in ultra-pure He atmosphere at 40 mL·min-1and trapped by a cold trap.The trap was then heated to 350 °C for 3 min,whereupon the analytes were vaporized and transferred to a gas chromatography-mass spectrometry (GC/MS,7890B/5977A,Agilent,USA).A DB-5 capillary column (30 m × 0.25 mm × 0.25 μm) was used.The initial column temperature was 150 °C,then increased to 250 °C at 10 °C·min-1,and kept for 2 min.The MS conditions and the quantification method were same as those of gas samples.As displayed in Table S2,a total of 54 species were analyzed from solid samples.

For the quality control(QC)and quality assurance(QA),laboratory blanks and field blanks were analyzed,and there was no significant contamination.For the two gas sampling methods,most of the recovery efficiencies of VOCs was in the range of 70% to 130%.Detection limits ranged from 0.05 to 0.1 mm3·m-3.

3.Results and Discussion

3.1.Emission characteristics of VOCs

3.1.1.Total mass concentrations

The flue gas was directly emitted into the surrounding air through the chimney after passing through sampling site #4(WFGD outlet) in the tested coal-fired power plant.Therefore,the VOC emission characteristics could be obtained directly according to the results of this site.The total mass concentration of the analyzed 102 species measured at sampling site #4 were(13456 ± 47) μg·m-3(collected by sampling bags) and(4166 ± 1034) μg·m-3(collected by SUMMA canisters),respectively.The results are in the same order of magnitude.However,the result obtained using the sampling bag was significantly higher and the relative deviation was smaller than those of the SUMMA canister.It indicates that when sampling VOCs in high humidity and high dust environment,the sampling bag sampling system may be more reliable than the SUMMA canister sampling system.

The total mass concentration of VOCs in this study was higher than the previous references,but was comparable to the most recent references.Fig.2 compares the concentration of several VOCs in these references and this study.It can be seen that the concentrations of benzene and toluene obtained with sampling bags in this study were close to those of Shiet al.[21],and much higher than those reported in other literatures.For species such as ethylbenzene,m/p-xylene,o-xylene,styrene,propylbenzene and trimethylbenzene,the highest concentrations were all in this study.The differences may be related to the number of VOC species tested and the sampling methods.For example,Garcia et al.[23]only analyzed 25 VOC species.Fernández-Martínez et al.[24,25]studied 16 VOC species.And dos Santoset al.[26] studied 36 species.The number of VOC species measured in the above studies was much smaller than the number in this study.In these studies,sorbent tubes were commonly used to collect gas samples.However,the adsorbents in the sorbent tubes have certain selectivity for organic matter.Therefore,the types of VOCs applicable to this method are limited,and the measurement results may not accurately reflect the total emissions of VOCs.Besides,the reasons for the differences in VOC emissions may also be correlated to the type of the coal used,the load of the power plant,the combustion conditions and the flue gas treatment methods.For instance,Garciaet al.[23]measured VOC emissions from a power plant with a load of 130 to 245 MW using a Polish-Colombian hybrid coal.The As Pontes power plant measured by Fernández-Martínezet al.[25]has an installed capacity of 350 MW and a mixture of subbituminous coal and lignite was used as fuel,while the Litoral power plant has an installed capacity of 550 MW and the coal burned was a South Africa bituminous coal.In Shiet al.’s [28]study,the power plant has a nominal power of 400 MW and the coal burned was a lignite with high volatile content (47.5%).In addition,power plants applying deep low NOxcombustion technologies in China may have lower local combustion efficiencies,and may also produce a significant amount of VOCs.It is not yet possible to determine the specific causes of the differences in VOC concentrations.Multiple sampling and comparisons can be carried out by controlling the above important variables in future studies.

Fig.2.Comparison of the emission concentration (μg·m-3) of several VOC species in published literature and this study (Note:a collected by sampling bags;b collected by SUMMA canisters;* load of the power plant (MW)).

On the other hand,it can also be seen from Fig.2 that sampling with SUMMA canisters may lead to the loss of data such as benzene and ethylbenzene.Since these organic components may occupy a considerable portion of the VOC mass fraction,sampling with this method could result in a lower total mass concentration of VOCs.Therefore,when analyzing the concentration and composition of the gas phase VOCs in the subsequent sections,the data obtained by sampling in the sampling bags will be mainly used.

3.1.2.Mass distributions

To obtain the emission source profiles,the fingerprint information of the emission source is more important than the total emissions.Thus,Fig.3 shows the mass fractions of various VOCs at sampling site #4.As can be seen from this figure,the identified VOC species were classified inton-alkanes,branched alkanes,cycloalkanes,alkenes,alkynes,aromatic hydrocarbons,oxygencontaining compounds and halogenated hydrocarbons.Among them,aromatic hydrocarbons had the highest mass fractions(26.8%),which is consistent with the results of most studies[26,28].The following species were alkenes (24.7%),n-alkanes(18.6%),halogenated hydrocarbons (14.5%),branched alkanes(13.4%) and cycloalkanes (1.2%).The mass fractions of alkynes and oxygen-containing compounds were small (less than 1%).The results might be related to the combined effects of the original chemical structure of coal,combustion conditions,flue gas treatment equipment,and temperature changes along the process.

Fig.3.Mass fractions of various VOCs at sampling site #4.

Fig.4 shows the mass concentration of each VOCs at the same sampling site (#4).There were 1–2 main components innalkanes,branched alkanes and alkenes,and their mass concentrations were significantly higher than other components of the same type.For example,amongn-alkanes,n-hexane(#5) held the highest mass concentration,and 2-methylpentane(#16)and 1-hexene(#41) were the most abundant species of branched alkanes and alkenes,respectively.The aromatic hydrocarbon components were more uniformly distributed.There were up to 12 species with a mass fraction higher than 1%,of which toluene (#44) dominated(5.6%).The halogenated hydrocarbons were also abundant.There were up to 5 species with a mass fraction higher than 1%,and trichloromethane (#77) had the largest contribution to the total VOCs,accounting for 5.1%.

Fig.4.Mass concentration of each VOC species at sampling site #4 (see Table S1 for the species name corresponding to the ID number).

Fig.5.Mass fractions of various VOCs at all sampling sites (#1,#2,#3 and #4).

The top ten VOC species measured at sampling site#4 and their mass fractions were listed in Table 3,along with the literature data.In this study,the top ten VOC species had a total mass fraction of 73.5%.The species with a higher mass fraction were mainly hydrocarbons with a carbon number of 6 or higher(such as 1-hexene,nhexane,2-methylpentane and toluene)and halogenated hydrocarbons with a lower carbon number (such as trichloromethane and chloroethane).The possible reason for the higher mass fractions of hydrocarbons with a higher carbon number may be that they were not easily oxidized during the reaction.While halogenated hydrocarbons can be released directly from the coal,or can be further produced by substitution reactions during the coal combustion process [24].The toxic species were mainlyn-hexane,toluene,m/p-xylene ando-xylene.Comparing our results with the results of Shiet al.[28],it can be found that half of the top ten VOC species were identical,and three of the first four species were the same.However,there was a clear different between the results of Yanet al.and our results [29].For example,1-hexene,the main species,did not appear in the results of Yanet al.[29].

3.2.Effects of the APCDs(SCR,FF and WFGD)on the transformation of VOCs

3.2.1.Mass concentrations of VOCs in the gas phase

Table 4 and Fig.5 display the total mass concentration of VOCs collected by sampling bags and the mass fractions of different types of VOCs measured at sampling sites #1,#2,#3 and #4,respectively.By comparing the total mass concentration of gas phase VOCs at each sampling site(see Table 4),it can be seen that after the pollutant control devices(SCR,followed by FF and WFGD),the total mass concentration of the 102 species analyzed slightly increased from (12399 ± 1265) to (13456 ± 47) μg·m-3.The mass distribution of different types of gas phase VOCs at each sampling site (see Fig.5) also changed slightly.The mass fraction of alkanes(includingn-alkanes,branched alkanes and cycloalkanes),aromatic hydrocarbons and halogenated hydrocarbons showed a downward trend,while the mass fraction of alkenes showed an upward trend.For example,after SCR,the percentage of branched alkanes felled from 18.0% to 14.1%,aromatic hydrocarbons decreased from 29.0% to 24.2%,and alkenes increased from 12.4%to 19.3%.The phenomenon indicated that SCR catalyst had selective oxidation for VOCs.After the dust removal device (FF),halogenated hydrocarbons were reduced by 5.5%,and alkenes were further increased by 3.4%.The proportion of each component did not change much before and after WFGD,all within 2.5%.

Table 5 shows the variation of the mass fractions of the characteristic VOCs (top ten VOCs of the WFGD exit).Except for the absence of chloroethane at the first three sampling sites,the massfractions of other characteristic VOCs at the four sampling sites all exceeded 1.0%,which means that most of the pollutants were mainly formed in the furnace.In addition,after each pollutant control device,the mass fractions of 1-hexene,n-hexane and toluene increased significantly,which may be caused by the drop of flue gas temperature or the internal physical/chemical processes in the pollutant removal devices.

Table 3Top ten VOC species at sampling site #4 with their mass fractions (comparing to literature data)

Table 4Total mass concentrations of VOCs at all sampling sites (#1,#2,#3 and #4)

3.2.2.Mass concentrations of VOCs in the particle phase

For VOCs existed in the particle phase,sampling and analysis of solid samples were also carried out.The mass concentration (in μg·g-1) of each VOC species in the solid samples collected at sampling sites#1,#2 and#4 was shown in Fig.6.Since the aperture of the sampling hole at the WFGD entrance (sampling site #3) is too small,data cannot be acquired at this sampling site.According to our measurements,the total mass concentrations of the analyzed 54 species measured at sampling site #1,#2 and #4 were 1.8,3.7 and 100.3 μg·g-1respectively.The reasons for the sharp increase in the mass concentration may be as follows:on one hand,as the flue gas temperature decreases and the oxygen content increases,the condensation of gas phase VOCs (homogeneous nucleation),the adsorption of gas phase VOCs on the particle surface,and the absorption of gas phase VOCs into the particles(heterogeneous nucleation) may occur,which in turn affects the gas/particle distribution of VOCs[31].These processes involve both physical (condensation,etc.) and chemical processes (oxidation and heterogeneous reactions,etc.),which all lead to the increase of the mass concentration of VOCs in the particle phase.Since existing studies have revealed that heterogeneous nucleation has crucial effect on determining gas/particle partitioning,this effect is referred to herein as the adsorption/absorption effect.On the other hand,due to the wide distribution of particle size of the particulate matter in the flue gas and the low removal rate of dust removal device for particles ranging from 0.1 to 1 μm [32,33],the particles at the FF outlet are generally small in particle size and large in surface area.These particles have a strong ability to adsorb gas phase VOCs,thus the mass concentration of VOCs in the solid samples after the FF is much higher.This effect is referred to herein as the particle size screening effect.The main VOCs in the particle phase of the three sites were basically the same,and the most important species was 2,2-dichloropropane (No.33).A considerable concentration of 1,2,3-trichloropropane (No.34) and tetrachloroethylene (No.30) were also measured at sampling sites #1 and #2.

Fig.6.Mass concentration of each VOC species in the solid samples collected at all sampling sites.(See Table S2 for the species name corresponding to the ID number).

In addition,the total mass concentrations(in μg·m-3)of VOCs in the particle phase at the three sites could be calculated according to the dust concentrations in the flue gas(see Table 1),which were 36.6,45.5 and 0.157 μg·m-3,respectively.The results showed that FF and WFGD are quite effective,while the effect of SCR is not obvious,accompanied by a slight increase in the total mass concentration of VOCs in the particle phase.The total mass concentration of VOCs in the particle phase was greatly reduced due to the decrease in dust concentration.Therefore,improving the efficiency of the dust removal and desulfurization equipment,especially the removal efficiency of fine particles,could effectively reduce the total amount of particle phase VOCs discharged into the atmosphere.However,since the mass concentration of VOCs in the particle phase was significantly lower than that in the gas phase,the synergistic effect of the pollutant control devices on the total VOC emissions is limited.Thus,it’s necessary to develop control technologies to reduce the emission of VOCs in power plants.Promising control technologies mainly include improved fly ash injection technology [30] and the modification of SCR catalyst to remove VOCs simultaneously [34].

Table 5Top ten VOC species at all sampling sites with their mass fractions (%)

3.2.3.Migration behavior of VOCs

To further analyze the effect of the pollutant control devices on the migration behavior of VOCs,as displayed in Fig.7,the flow dia-gram of VOCs was obtained under the assumption that the substances (i.e.,VOCs) were balanced.During the SCR process,there was no significant change in the total mass flow rate of VOCs(with a reduction of 4.17%).It indicates that the V-W/Ti catalysts had a limited influence on VOCs removal in this process.According to the existing literatures,V2O5/TiO2-based catalysts could catalyze 1,2-dichlorobenzene [35];after adding WO3,the V2O5-WO3-TiO2catalysts can catalyze propane,propene,isopropanol,acetone,1-/2-chloropropane and 1,2-dichlorobenzene when excess oxygen existed.Leeet al.’s[36]bench-scale experiments showed that commercial SCR catalyst could synergistically remove benzene and NOxin simulated flue gas conditions.Finocchioet al.[37] reported VW-Ti catalysts with high V content were more active.The V-W-Ti catalysts showed different activities under different flue gas conditions.For example,the catalysts were more active to remove halogenated hydrocarbons than hydrocarbons,especially in the dehydrochlorination of the chloropropane isomers,and could maintain the same activity when HCl existed [37,38].Besides,the catalytic activity of the V-W-Ti catalysts was also affected by the H2O concentration [39] and the NO concentration [40,41].Thus,the low removal efficiency of SCR process may be owing to the comprehensive influence of the above-mentioned various factors.

Fig.7.Flow diagram of VOCs in the flue gas from the tested coal-fired power station unit.(the numbers outside and inside the brackets represent the mass flow rate of total VOCs and gas phase VOCs,respectively,in kg VOCs·h-1).

During the dust removal process,the total mass flow rate of VOCs was increased by 28.70%,although VOCs in the particles phase were captured (the mass flow rate of particle phase VOCs was reduced by 3.48%).The increase may be related to the following three reasons:(1) NOx,SOxand the halogenated hydrocarbons in the flue gas might react with the existing VOCs,forming new VOC components;(2) Polytetrafluoroethylene (PTFE) based cloth and polyphenylene sulfide fiber+polytetrafluoroethylene(PPS+PTFE)blended filter material used in the FF may decompose at the operating temperature (120–140 ℃) and release VOC components;(3) The connecting parts of the air preheater and the FF,the filter material and the captured fly ash may adsorb/absorb some VOCs and then release them under certain conditions.

Finally,in the desulfurization process,the total mass flow rate of VOCs was reduced by 18.24%.This indicated that WFGD was the most effective device on the reduction of VOCs among the three conventional pollutant control devices studied in this paper.In WFGD,the washing and trapping action of the desulfurization slurry might play an important role.This study did not discuss the above analysis in depth because further research is needed.

3.3.Ozone formation potential (OFP)

In order to evaluate the chemical reactivity of individual VOCs,the ozone formation potentials(OFP)of each VOC species was calculated by the following equation [42]

where OFPirepresents the OFP of theith species (μg ozone·m-3),MIRirepresents the maximum increment reactivity of theith species (μg ozone·μg-1),Cirepresents the mass concentration of theith species (μg·m-3).The total OFP could be obtained by summing all OFPs of each VOC species.

Through the use of MIR values in literature,the total OFPs(～60 species)at the four sampling sites were calculated to be 37.1,36.3,51.6 and 46.4 mg ozone·m-3,respectively.These values inferred that the total OFP increased slightly through the conventional pollutant control devices,which was mainly caused by the dust removal process.

The top ten VOCs that contributed the most to ozone formation were shown in Fig.8 with their contributions to OFP(%).As can be seen from this figure,the top ten species showed little change at the four different sampling sites.At the WFGD exit,1-hexene contributed up to 39.1% to the total ozone formation due to its high mass concentration and high MIR value (5.49).n-hexane and 2-methylpentane were also listed in the top ten OPF contributors because of their relatively high mass concentrations (see Fig.4).However,their MIR values were not significant (1.24 and 1.5,respectively).Thus their contributions to OFP were much smaller than their contributions to the total mass concentrations (see Table 5).In addition to the above three VOC species,the contributions of aromatic hydrocarbons to the total ozone formation must also be considered.As shown in Fig.4,the mass concentrations of aromatic hydrocarbons ranked high among all VOCs tested (most of the top 20 species were aromatic hydrocarbons)and their corresponding MIR values were also high.For instance,the MIR value ofo-xylene is 7.64,the MIR value of 3-ethyltoluene is 7.39,and the MIR value of trimethylbenzene is between 8.87 and 11.97.Therefore,the contribution of aromatic hydrocarbons to OFP reached 46.4%,which is the highest among all kinds of VOCs.While halogenated hydrocarbons,such as trichloromethane and chloroethane,had a low contribution to OFP due to their extremely low MIR values (0.022 and 0.29,respectively).In general,we can see that the contribution of a certain VOC species or a certain type of VOC species to OFP depends not only on the mass concentration of the species,but also on its MIR values.

Fig.8.Top ten VOCs species at all sampling sites (#1,#2,#3 and #4) with their OFP contributions.

4.Conclusions

A study about the emissions characteristics of VOCs from a coalfired power plant was carried out by analyzing the VOC mass concentrations in both gas and solid samples using GC/FID/MS and GC/MS.The effects of the existing conventional pollutant control devices (SCR,FF and WFGD) on the transformation of VOCs were explored,and the influence of the VOC species on ozone formation was evaluated.We found that the sampling bag sampling system may be more reliable than the SUMMA canister.The total mass concentration of the measured 102 species collected by sampling bags in the WFGD outlet was (13456 ± 47) μg·m-3.Among them,aliphatic hydrocarbons had the highest mass fractions(57.9%),followed by aromatic hydrocarbons (26.8%) and halogenated hydrocarbons (14.5%).The top ten VOC species in terms of mass fraction were mainly hydrocarbons with a carbon number of 6 or higher(such as 1-hexene,n-hexane,2-methylpentane and toluene)and halogenated hydrocarbons with a lower carbon number (such as trichloromethane and chloroethane).Conventional pollutant control equipment (SCR,FF and WFGD) had a limited effect on VOC reduction.The FF process even caused the increase of gas phase VOCs.WFGD was the most effective device on the reduction of VOCs among the three conventional pollutant control devices in this study.OFP calculations showed that aromatic hydrocarbons contributed the highest ozone formation (46.4%) owing to the relatively high mass concentrations and MIR values.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was funded by the National Natural Science Foundation of China(52006079),the Natural Science Foundation of Hubei Province (2020CFB247) and the National Key Research and Development Program of China (2018YFB0605201).

Supplementary Material

VOC species measured in gas and solid samples.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.02.015.