Characterization of PAHs and PCBs in Fly Ashes of Eighteen Coal-Fired Power Plants

For the management of coal fly ashes (CFAs) from coal-fired power plants (CFPPs), characterization of PAHs and PCBs in CFAs is imperative. The 18 PAH and 86 PCB congeners in CFAs collected from 18 large-scale CFPPs in China were detected using GC/MS system. The PAH concentrations were in the range of 5.51–70.9 ng g for 16 CFPPs with individual block power capacity as 600 MW (IBPC-600), significantly lower than 886–916 ng g for 2 CFPPs with IBPC as 200 and 300 MW (IBPC-200/300). Both PAH and PCB congeners for 18 CFPPs were dominated by low molecular weight ones. The 3and 2-ring PAHs, di-, triand tetra-PCBs were the predominant homologs. PAH profiles for 16 CFPPs with IBPC-600 were significantly different from other industrial stacks based on higher coefficients of divergence. The BaP-based toxic equivalency (BaPeq) concentration and BaP-based equivalent carcinogenic power (BaPE) for 16 CFPPs with IBPC-600 were 0.834 ng g and 0.570, much lower than corresponding 20.5 ng g and 15.4 for 2 CFPPs with IBPC200/300. No difference existed for Σ86PCBs between CPFFs with IBPC-600 and -200/300, which ranged from 9.60 to 32.1 ng g. Higher mean carcinogenic PAH concentrations for 2 CFPPs with IBPC-200/300 and PCBs-TEQ concentration for 18 CFPPs indicated the application of CFAs as soil amendment should be prohibited. The PAH concentrations for 18 CFPPs were well correlated with the total organic carbon (TOC) values, while PCB concentrations showed not this trend, indicated the different formation mechanism between PCBs and PAHs.


INTRODUCTION
China is the world-largest coal producer and consumer, which consumed 50.2% of the world coal in 2012 (Yao et al., 2015).The huge quantity of CFAs produced from burning process of pulverized coal in CFPPs can be a problematic solid waste (Skousen et al., 2013;Ram and Masto, 2014).CFA would be recognized as environmental pollutant if not put to beneficial use (Blissett and Rowson, 2012).In China, CFPPs accounted for 78% of the total power generation in 2012.Pandey and Singh (2010) reported as many as 100 million tons were generated in China for 2005 and only 45% of them were utilized.In China the same installed capacity as those for the whole 20 th century would be completed in next 20 years (Shaheen et al., 2014).Although the part of this increased demand would be met by renewable resources, coal is still the dominant fuel for power generation in China and India (Lior, 2010).Annual CFAs generation in China has been increased during 2001-2015 and is still increasing.The predicted 580 million tons should be reached in 2015 (Yao et al., 2014).Clearly the more and more CFAs must be disposed in future China.Some recycling methods of CFA have been reported in recently documents (Blissett and Rowson, 2012;Shaheen et al., 2014).The CFAs can be used as a soil amendment, catalysts and catalyst supports, or used in concrete and cement manufacturing.However, these applications only consume a small proportion of the CFAs.The rest more than half of CFAs have to be disposed in landfills (Sahu et al., 2009;Yao et al., 2015).The CFAs were often stockpiled in open area before landfill or storing in lagoons in the present China, and the large amounts of pollutants in CFAs could easily entered into different environmental medias.Higher unintentionally produced persistent organic pollutants (UP-POPs) levels have been found in CFAs from CFPPs (Ba et al., 2009).The UP-POPs contained in CFAs could contaminate the soil, air and ground water if not treated properly (Yao et al., 2015).The work burden of CFA disposal is becoming heavier due to its steadily risen production and uncertain properties in China.For the large-scale environmental safe use and disposal of CFAs, the detailed study on UP-POPs in CFAs is imperative (Sun et al., 2006;Wang et al., 2013).
PAHs are belong to POPs and have been a concern for their threats to people's health (Zhang et al., 2009;Kong et al., 2011;Wu et al., 2014;Tiwari et al., 2015;Huang et al., 2016;Pongpiachan et al., 2016), mainly originated from burning and pyrolytic processes (Chen et al., 2014;Li et al., 2014).Coal burning for power generation has been considered one of the most significant anthropogenic source of PAHs by some researchers (Mastral and Callén, 2000;Liu et al., 2002;Arditsoglou et al., 2004;Sahu et al., 2009).PCBs are a kind of classic POPs due to their high lipophilicity, toxicity and difficulty of degradation (Guo et al., 2014;Liu et al., 2014).Although CFPPs is a known PCBs emission sources (Jansson et al., 2011;Li et al., 2011Li et al., , 2015)), the PCBs emission characteristics in CFAs from CFPPs has not been studied extensively as other POPs in China.The significant difference of PCB profiles existed among the flue gases, CFAs and Aroclor mixtures, possibly resulted from different formation mechanisms, so the systematically research on PCBs in CFAs is needed (Jansson et al., 2011).Sahu et al. (2009) reported the PAHs and PCBs levels in CFAs of CFPPs based on 5 Indian power plants.However, the data are not available for CFPPs in present China.
In present China, the CFPPs with lower IBPC are planned replaced by higher IBPC as 600 or 1000 MW in order to work economically and protect environment effectively.IBPC-600 is becoming the mainstream unit in China and account for 39% of present CFPPs and this ratio will increase in future (Zhu et al., 2014).The main aims of this study are to systematically investigate the PAHs and PCBs in CFAs from CFPPs with different IBPC in China.

CFA Sample Collection and Preparation before Extraction
As the Fig. 1, the sampling locations for 18 CFPPs covered essentially the whole Chinese territory.Along with the rapid economic development and strict requirement for environmental protection in China, higher IBPC is becoming the mainstream generation units for Chinese CFPPs.In present China, IBPC-600 is the mainly representative generation unit and accounts 39% of all the coal-fired power generation units.The samples from 16 CFPPs with IBPC-600 were FA1-4 and FA7-18.The CFA samples of FA5 and FA6 were collected from 2 CFPPs with IBPC-200 and -300, respectively.
The CFA samples were collected systematically from 18 large-scaled CFPPs in China during January to October, 2015.The same systematic sampling method as Sahu et al. (2009) was adopted in this study.The samples were collected using a stainless spoon from the bottom piles of the electrostatic precipitators and then stored in sealed polyethylene bags.About 1kg of sample for each CFPP was stored in amber glass jars pre-cleaned using solvents.Each sample was dried using a vacuum freeze dryer and weighed.

Sample Extraction for PAHs Analysis
The sample extraction and analysis procedures were used according to the EPA method TO-13A and detailed described in Kong et al. (2011).The 30 g fly ashes were ultrasonically extracted with dichloromethane, concentrated using a rotary evaporator, purified using a silica gel column and re-concentrated using a rotary evaporator.The HP6890 GC/5973i MS system with selected ion mode (SIM) was used to PAHs analysis.The chromatographic condition, temperature climbing program, PAH identification and quantification method were all same as Kong et al. (2011).GC/MS system was processed using selected ion monitoring (SIM) mode with m/z as 129, 127 for NaP, 153, 152 for Acy,151,153 for Ace,165,167 for Fl,179,176 for Phe and Ant,101,203 for Flu and Pyr,229,226 for BaA,226,229 for Chr,256,126 for BbF and BkF,253,126 for BaP and BeP,138,227 for Ind,139,279 for DBA,138,227 for BghiP,150,301 for COR, respectively.The m/z values as 128,154,152,166,178,178,202,202,228,228,252,252,252,252,276,278,276,300 were selected to quantify the PAHs as NaP, Acy, Ace, Fl, Phe, Ant, Flu, Pyr, BaA, Chr, BbF, BkF, BaP, BeP, Ind, DBA, BghiP and COR, respectively.

Quality Control and Quality Assurance
For both PAH and PCB analyses, a procedural blank, a matrix-spiked sample, a sample duplicate and a solvent blank were processed every 5 samples.No any target compounds were detected in the procedural and solvent blank.
For PAH analysis, the method detection limits (MDLs) of 18 PAH congeners ranged from 0.05 for BaP to 0.60 ng g -1 for NaP with the mean value as 0.16 ng g -1 .The recoveries for 18 PAH congeners in 4 matrix-spiked samples were in the range of 79-115% and the relative standard deviation (RSD) was lower than 12%.The recoveries for surrogate standard of 14-deuterium substituted terphenyl and 4-bromo-2-fluorobiphenyl were 97.20 ± 12.22% and 89.36 ± 11.66%, respectively.The method detection limits for 18 PAH congeners were ranged from For PCB analysis, the recoveries for surrogate standard of PCB-14, 65, and 166 were 93 ± 15%, 89 ± 10 and 88 ± 13%, respectively.The recoveries for 86 PCB congeners in 4 matrix-spiked samples were all within 85-110% and the RSD was lower than 9%.The standard deviations for 4 pairs of repeated samples were all lower than 11%.The MDLs for 86 PCB congeners ranged from 2 to 33 pg g -1 with the mean values as 7.54 pg g -1 .

PAHs Concentrations in CFA Samples for 18 of Chinese CFPPs
The total amounts of 18 PAH congeners were ranged from 5.51 to 70.9 ng g -1 (mean: 26.94 ng g -1 ) for CFPPs with IBPC-600, and ranged from 886 to 916 ng g -1 (mean: 901 ng g -1 ) for CFPPs with IBPC-200/300 (Table 1).The results indicated the influence of different combustion condition, dust removal technique, the burn-up of feed coal among CFPPs with different IBPC values (Mastral et al., 2000;Yang et al., 2002;Kong et al., 2011;Masto et al., 2015).Sloss and Smith (1993) reported three formation mechanisms for PAHs during combustion process: 1) the penetration of feed coal containing-PAHs through combustion area without being destroyed; 2) reformation of ring structures; 3) de novo synthesis of PAHs via oligomerisation and cyclisation of acetylenes formed as a consequence of incomplete combustion.
The combustion temperature of coal among the 18 CFPPs were similar, the TOC concentration was the main reason result in this difference.The TOC and PAH concentrations for 18 CFPPs were well correlated (R 2 = 0.93, p < 0.005), indicated that the incompletely combustion of coal is the main formation mechanism of PAH in CFAs from CFPPs (Fig. 2).Wang et al. (2013) also reported they were well correlated (R 2 = 0.96) for PAH congeners in fly and bottom ashes of a Huainan CFPP in China.The unburned organic carbon was known as a stronger sorbent for PAHs than minerals (Karapanagioti et al., 2000).The operating conditions of  CFPPs such as combustion efficiencies and excess oxygen were the key factors controlling PAH emissions, more important than type of coal and plant design (Revuelta et al., 1999;Cheruiyot et al., 2015).Significant variation existed in PAH compositions among the different combustion processes (Yang et al., 2002;Kong et al., 2011).Sohu et al. (2009) reported the mean value of Σ 14 PAHs for 5 Indian CFPPs was 355 ng g -1 , which was compared to that for FA5 (886 ng g -1 ) and FA6 (917 ng g -1 ) in this study, much higher than mean value for 18 CFPPs with IBPC-600 (26.9 ng g -1 ).The difference among IBPC values, combustion temperature, combustion technique and feed coal would be the explanation.The 26.9 ng g -1 for IBPC-600 was similar to the 32.1 ng g -1 for a CFPP with IBPC-600 located at Unchahar, India (Verma et al., 2015) and 9.78 ng g -1 for a CFPP with IBPC-620 in Belgrade (Pergal et al., 2014).
As for CFPPs with IBPC-200, the 916 ng g -1 was lower than that (766 µg g -1 ) of PM 10 fraction from a CFPP with IBPC-200 for heating in Liaoning province, China (Kong et al., 2011), possibly resulted from the shorter residence time and larger surface of PM 10 .Arditsoglou et al. (2004) and Wang et al. (2015) also reported the reverse relationship between PAH levels and particle size of CFAs.Li et al. (2014) also reported higher PAH concentrations in fine particles were related to their long residence time in combustion area and high surface area.
The PAH levels of CFPPs were significantly lower than those of biomass fired power plants (BFPPs).Masto et al. (2015) reported the concentrations of total 16 PAH congeners were 0.19-12.3mg kg -1 in fly ashes from 4 Indian BFPPs, which suggested the influence of IBPC values and biomass fuels on PAH concentrations.
For the individual PAH congener (IPC), Nap, Phe, Fl and Flu were the predominant congeners in CFAs from 18 CFPPs and accounted for 69.9% of Σ 18 PAHs (Table 1).
This IPC distribution possibly resulted from the different temperature of flue gas at the different part of boiler.For example, the flue gas temperature were 500°C for coal saving device, 800-900°C for superheater, and 1200-1400°C for boiler chamber, respectively.The highest concentration for each IPC appeared at a specific temperature.For examples, Ace at 700°C, NaP, Flu, Chr and BaP at 800°C and the others at 900°C.Ant, Phe and Flu increased when the temperature rise from 800 to 900°C (Peng et al., 2016).Peng et al. (2016) reported the top congeners were Phe and Ant during the coal combustion process.Sahu et al. (2009) reported they were Fl, NaP, BaA, DBA and Flu for 5 Indian power plants.Kong et al. (2011) reported they were Ant, Flu and Phe for PM 10 collected from FA of a coal-fired power plant in Liaoning aim to provide heating for local residents using a re-suspension sampler.Singh et al. (2013) reported they were Ant, Flu, Pyr and BaA in the fly ash from combustion of rural used biomass fuel in Indo-Gangetic plains of India.The difference of IPC profiles possibly result from the combustion condition (such as turbulence, temperatures and residence time), fuel composition, and air flow (Sloss and Smith, 1993;Cheruiyot et al, 2015;Košnář et al., 2016;Peng et al., 2016;Wen et al., 2016).

Potential Toxicity Risk Assessment for PAHs
The values of BaP-based toxic equivalency concentration (BaPeq), total carcinogenic PAHs (C-PAHs), BaP-based equivalent carcinogenic power (BaPE) and 2,3,7,8-tetrachlorodibenzodioxin (TCDD)-based toxic equivalency concentration (TEQ), are often calculated and used to evaluate the potential health risk of PAHs to human beings in previous studies (Chen et al., 2004(Chen et al., , 2007;;Liu et al., 2009;Kong et al., 2011Kong et al., , 2013;;Cheruiyot et al, 2015).The calculated results of potential toxicity risk of PAHs in this study were shown in Table 2.The 16 CFPPs with IBPC-600 have a lower mean BaPeq as 0.834 ng g -1 than 20.5 ng g -1 for 2 CFPPs with IBPC-200/300 due to higher PAH levels for CFPPs with IBPC-200/300.The significantly difference among ratios of BaPeq of each PAH to that of BaP existed between CPFFs with IBPC-600 and -300/200, resulted from the difference of IPC concentrations and their BaP-based equivalent factors.DBA and BbF had the highest ratios as 77.5%, 46.6% for CPFFs with IBPC-600, while the ratios of BbF and Flu were 43.7% and 26.7% for CFPPs with IBPC-200/300.The total BaPE were 0.570 and 15.4 for CFPPs with IBPC-600 and -200/300, respectively.The incompletely combustion of coal in CFPPs with IBPC-200/300 possibly resulted in these difference.
As an indicator PAH, BaP was the most carcinogenic congener (Kong et al., 2011).Its contents in this study ranged from 0.170 to 0.710 ng g -1 for CFPPs with IBPC-600, 1.33 to 17.46 ng g -1 for CPFFs with IBPC-200/300, respectively.According to the commonly used PAH criterions for sensitive land use designated by the Swedish EPA as 0.3 mg kg -1 for carcinogenic PAHs (Sun et al., 2006;Zhao et al., 2008;Kong et al., 2011), the 3.88 ng g -1 of carcinogenic PAHs for CFPPs as IBPC-600 was significantly lower than 0.3 mg kg -1 , while the corresponding value of 0.11 mg kg -1 for CFPPs with IBPC-200/300 was comparable to 0.3 mg kg -1 .Therefore, the utilization of CFAs as soil amendment from CFPPs with IBPC-200/300 should be prohibited (Table 2).

Source Profiles of PAHs
To our knowledge, the data of PAH source profiles in CFAs from CFPPs across China were not available.Similarities between the PAH profiles in fly ashes from different industrial stacks were compared based on coefficient of divergence by previous documents, a self normalizing parameter to check the data points expansion for two data groups.The parameter was calculated as follows.
where j and k are referred to the two profiles for different emission sources, p was the number of detected PAH congeners, and x ij and x ik were the mean mass concentrations of compound i for j and k (Wongphatarakul et al., 1998;Kong et al., 2011).d The BaP-equivalent carcinogenic power (BaPE) for the total PAHs: BaPE = BaA × 0.06 + B[b,k]F × 0.07 + BaP + DBA × 0.6 + Ind × 0.08 (Chen et al., 2004;Liu et al., 2009;Kong et al., 2011).
Source profiles of j and k were similar if the CD jk approaches zero and they were significantly different if it approaches one (Wongphatarakul et al., 1998).Wongphatarakul and Kong thought the CD of 0.269 and < 0.3 suggested the similarity between two source profiles in their studies (Wongphatarakul et al., 1998;Kong et al., 2011).
In this study, the CD jk were calculated for CFAs from 16 CFPPs with IBPC-600 and fly ashes from other industrial stacks (Sahu et al., 2009;Kong et al., 2011).They were 0.645 for CFPPs vs. coke production, 0.674 for CFPPs vs. iron smelting process, 0.587 for CFPPs vs. 5 India CFPPs and 0.608 for CFPPs vs. BFPPs, respectively.These higher CD jk values indicated the significantly difference existed between the Chinese CFPPs with mainstream IBPC and other industrial stacks.As shown in Fig. 4, the difference among the mass concentrations of Nap, Phe and Fl possibly resulted in higher CD jk values.

PAH Diagnostic Ratios
The diagnostic ratios (DRs) were often used to identify the PAH emission sources and the use and selection of DRs in PAH source apportionment was described elsewhere (Ravindra et al., 2008;Tobiszewski and Namiesnik, 2012).The DRs were used to replace the PAHs profiles for emission sources although the PAHs profiles were always different among PAHs sources, to eliminate the effect of chemical reactions with other atmospheric pollutants.The DRs widely used to investigate PAH origins were calculated and listed in Table 3.
Seven indicator PCB congeners (IPCBs) including 52,101,118,138,153,and 180 were commonly measured and often compared in PCBs analysis (Li et al., 2011).In this study, the ΣIPCBs ranged from 1.72 for to 6.09 ng g -1 with a mean value as 2.75 ng g -1 , which were weak correlated with total PCBs (R 2 = 0.16).The lower correlation coefficient was possibly attributed to the formation mechanism of PCBs in coal combustion process, while the IPCBs were mainly contained in the conventional PCB products.Li et al. (2011) also reported this lower coefficient for atmospheric PCBs in east Tianjin due to the PCB emissions from new sources.The 11 dioxin-like PCBs (dl-PCBs) except PCB-77 such as  were all detected in this study, and the TEQ was calculated for each dl-PCB.The toxic equivalency (TEQ) concentration was calculated for each dl-PCB through multiplying its concentration by its WHO 2005 toxic  equivalent factor (TEF) for human and mammals (Li et al., 2011(Li et al., , 2015)).The sum of 11 dl-PCBs ranged from 2.23 ng g -1 for FA8 to 6.52 ng g -1 for FA5 with the mean value as 3.35 ng g -1 , which were in the range of 0.00-13.8ng g -1 for dl-PCBs in fly ashes from municipal solid waste incinerators (MSWI) in Korea reported by Shin and Chang (1999), higher than 1.65 ng g -1 for 15 large-scale MSWIs in China reported by Pan et al. (2013) and 0.52-2.3ng g -1 for 4 MSWIs in China reported by Wu et al. (2016).While Cobo et al. (2009) reported the concentration of dl-PCBs as high as 49.0 ng g -1 for FA from a MSWI in Medellin, Colombia.These differences for dl-PCB concentrations were possibly resulted from the different combustion condition and fuel components among different incinerators.As shown in Fig. 6(a), the top dl-like PCBs were PCB-169, -189, -118 and -157, the corresponding concentrations were 1.05, 0.61, 0.40 and 0.30 ng g -1 , respectively.As shown in Fig. 6(b), PCB-169 and -126 have the higher TEQ concentrations as 31.5 and 21.1 pg g -1 due to their high concentration and TEF values, which contributed 59.8% and 40.1% to the total TEQ concentration.The TEQ concentrations for 18 CFPPs in the range of 38.8-78.6 pg g -1 (mean: 52.6 pg g -1 ), significantly higher than 4 pg g -1 designated by Canada for soil quality (CCME, 2007), which did not supports their suitability for soil amendment.

CONCLUSIONS
The 18 PAH and 86 PCB congeners in CFAs collected from 18 large-scale CFPPs across China were detected using GC/MS system.The Ʃ18PAHs ranged from 5.51 to 70.9 ng g -1 with the mean value as 26.9 ng g -1 for 16 CFPPs with IBPC-600 and from 886 to 916 ng g -1 with the mean value as 901 ng g -1 for 2 CFPPs with IBPC-200/300.The PAHs in CFAs from 18 CFPPs were dominated by 3-and 2-ring congeners, which accounted for 40.5% and 29.6%, respectively.For the individual PAH, Nap, Phe, Fl and Flu were the predominant congeners in CFAs from Chinese 18 CFPPs and accounted for 69.9% of Σ 18 PAHs.PAH profiles for 16 CFPPs with IBPC-600 were significantly different from the other industrial stacks such as coke production, iron smelting process, Indian 5 CFPPs and 4 BFPPs with higher coefficients of divergence as 0.645, 0.674, 0.587 and 0.608, respectively.The total BaP-based toxic equivalency (BaPeq) concentration and BaP-based equivalent carcinogenic power (BaPE) for CFPPs with IBPC-600 were 0.834 ng g -1 and 0.570, lower than corresponding mean values of 20.5 ng g -1 and 15.4 for 2 CFPPs with IBPC-200/300.The diagnostic ratios (DRs) of CFPPs with IBPC-600 and -200/300 were generally similar.The DRs of Ant/(Ant + Phe), Flu/(Flu + Pyr), Ind/(Ind + BghiP), BaP/BghiP, BaA/(BaA + Chr) in this study were all similar with the corresponding reported values for coal burning.The DR values for different industrial stacks varied significantly each other, such as BaP/BghiP, Pyr/BaP and BaA/BaP for Chinese 18 CFPPs were significantly different from Indian 5 CFPPs.No significant difference existed for total PCB concentrations between CFPPs with IBPC-600 and -200/300.The sum of 86 PCB congeners for 18 CFPPs ranged from 9.60 ng g -1 for FA10 to 32.1 ng g -1 for FA11 with the mean value as 16.0 ng g -1 .PCB-4/10, -28, -6, -18, -169, -52 were the predominant congeners for 18 CFPPs.The CFAs from 18 CFPPs were dominated by LMW-PCBs with di-, tri-and tetra-PCBs contributed 28.7%, 35.6% and 10.2% to Ʃ 86 PCBs.PCB-169 and -126 have the highest TEQ concentrations as 31.5 and 21.1 pg g -1 , respectively.The PAHs concentrations were well correlated with the TOC (R 2 = 0.934, p < 0.005) and PCBs showed not this trend (R 2 = 0.063, p < 0.005), indicated the different formation mechanism between PCBs and PAHs.
Higher carcinogenic PAH levels for CFPPs with IBPC-200/300 and PCB TEQ concentrations for 18 CFPPs indicated the using of CFAs as soil amendment should be prohibited.

Fig. 2 .
Fig. 2. PAH concentrations and TOC values for each CFA sample.

Fig. 3 .
Fig. 3. Composition profiles of PAHs by ring size for a) individual fly ash samples and b) all the fly ash samples

Fig. 5 .
Fig. 5. (a) Concentrations of PCBs with different chlorine atoms and total PCBs in CFA from each CFPP and (b) PCB homologs distribution in CFA from 18 CFPPs.

Table 2 .
BaPeq and BaPE for CFPPs with different IBPC and BaPeq ratios for individual PAH congener.

Table 3 .
Diagnostic ratios for several PAHs for ashes from various industrial stacks.
a SD: Standard deviation.