Characterization of Inorganic Elements within PM 2 . 5 and PM 10 Fractions of Fly Ashes from Coal-Fired Power Plants

In this study 15 fly ash samples were collected from 15 large-scale coal-fired power plants (CFPPs) in China. The samples were then re-suspended through PM2.5 and PM10 inlets and analyzed for the contents of 39 inorganic elements (IEs) using inductively coupled plasma-mass spectroscopy (ICP-MS) and inductively coupled plasma-optical emission spectrometry (ICP-OES). The results show that the particle size distributions for the 15 FAs samples exhibited bimodal patterns. The Ʃ39IEs (g g) for the PM2.5 (0.292–0.564) in all the 15 CFPPs were higher than that of PM10 (0.269–0.403). Except for Cu, all the other 38 IEs were more enriched in the PM2.5 with the PM2.5/PM10 ratios being 1.06–1.73. Considering 13 heavy metals, the same orders occurred between PM2.5 and PM10 with Al >> Cr > Zn > Mn > Cu > V > Pb > Sn > Co > As > Sb > Tl > Cd. More attention should be paid to the high contents of Cr in both PM2.5 (1310 mg g) and for PM10 (1240 mg g) from all 15 CFPPs. 23 IEs for PM10 and 26 IEs for PM2.5 had the geo-accumulation index (Igeo) values higher than 0, indicating different pollution levels for them. On the other hand, there was moderate to extreme levels of pollution for Cr, Zn, Cu, Pb, Sn, Sb, Tl, Cd and Al based on Igeo values. The element profiles for PM2.5 or PM10 from 15 CFPPs were similar based on low coefficients of divergence for PM2.5 (0.254 ± 0.038) and PM10 (0.244 ± 0.054) according to the comparison between any two CFPPs. Most elements with low relative enrichment factors (REF) as less than 0.70 or 0.70–1.30 indicated no or weak condensation occurred for them during coal combustion, while Cr, Cu, Zn, Sn, W and Pb had REF values higher than 1.30 indicated that significant condensation occurred for these elements.


INTRODUCTION
Coal is still the most important energy source in the present China (Li et al., 2014).More than half of coal consumed in China is used for power generation and this amount is projected to increase in the foreseeable future (You and Xu, 2010).Fly ashes (FAs) account for 80% of the total coal burning by-products in pulverized coal fired power plants (CFPPs) (Janvijitsakul and Kuprianov, 2008).Large amounts of FAs originating from CFPPs are identified as problematic solid waste (Li et al., 2016).China is the largest FAs producer and it is predicted that the production of 570-610 million tons will be reached by the year 2020 (Cao et al., 2008;Li et al., 2014).Clearly more and more FAs must be effectively disposed in the future in China (Li et al., 2016).Although FAs have been widely used in many products and industries (Blissett and Rowson, 2012;Li et al., 2014;Shaheen et al., 2014;Li et al., 2016), the utilization rate is still low in China.The rest 55% of FAs are stockpiled before landfill (Sahu et al., 2009;Yao et al., 2015;Li et al., 2016).The reuse of FAs with different methods raises the concern about the characteristics of inorganic and organic pollutants in contained in the fly ashes (Baba et al., 2010;Li et al., 2014) including toxic compounds such as dioxins and PAHs (Cheruiyot et al., 2015;Cheruiyot et al., 2016).
Coal contains a large number of IEs that are subsequently emitted in the FAs or react with surface chemicals of FAs during the combustion process (Choi et al., 2002;Jegadeesan et al., 2008).Higher levels of element contents in FAs are 4-10 times those of the feed coal, which should be an environmental concern (Akar et al., 2012).The elements associated with stockpiled FAs may enter into soils and ground waters and result in adverse environment impacts (Li et al., 2016).Padhy et al. (2016) found that the toxic elements such as Pb, Cr and Cd in plants that were grown in FAs amended soils tended to increase significantly under the influence of elements in FAs.
Atmospheric particulate matters (PM), especially particles with lower aerodynamic diameter, have received significant attention in the present China (Jiang et al., 2015).The PM 2.5 and PM 10 within FAs can be easily re-suspended by wind during their storage in the landfills and result in the atmospheric PM pollution.These PM 2.5 and PM 10, can subsequently deposit on urban surfaces such as roads, soils, plants and building shells, then transported among environmental mediums by mixing force as wind, weathering, and gravitation and so on (Caravanos et al., 2006;Zhao et al., 2006;Shi et al., 2008).IEs associated with these particles can accumulate in human body via inhalation, ingestion and dermal contact hence causing acute and chronic toxic effects such as neuropathies, rise in blood pressure, lung cancer and other diseases (Kong et al., 2011;Li et al., 2013;Dai et al., 2015;Wei et al., 2016;Zheng et al., 2016).IEs associated with re-suspended PM 2.5 and PM 10 from ash ponds by wind have been of major concern due to their toxicity and long residence time in atmosphere (Kong et al., 2011;Li et al., 2014).Laidlaw and Filippelli (2008) and Laidlaw et al. (2012) found the re-suspended particles from soils had become a persistent source of blood Pb for children in many USA cities.
The correlation between particle size of FAs and element contents had been studied widely due to most of elements were enriched in finer particles and hard to remove from flue gas (Sarkar et al., 2006;Dai et al., 2010).To our knowledge, no detailed data of elements within fractions of PM 2.5 and PM 10 for FAs emitted from CFPPs was available (Xu et al., 1997;Wang et al., 1998;Li et al., 2014;Liu et al, 2015).So the study on inorganic elements abundant PM 2.5 and PM 10 within FAs would be a novel study.
In this study, 15 FA samples were systematically collected from 15 CFPPs in China with mainstream individual block of power capacity of 600 MW.One pared of PM 2.5 and PM 10 were sampled using re-suspension method for each FA for analysis of 39 IEs using ICP-MS and ICP-OES systems.
The main aims of this study were to investigate: 1) content distribution for sum of 39 IEs and individual element in both PM 2.5 and PM 10 ; 2) geo-accumulation index (Igeo) assessment for analysis of pollution levels of individual element in PM 2.5 and PM 10 ; 3) comparison of the similarity of the element composition profiles for both PM 2.5 or PM 10 between any two CFPPs; 4) enrichment factors for each element were determined in order to analyze the behavior of particular elements during coal combustion.

Sample Collection and Re-suspended for Obtaining of PM 2.5 and PM 10
The FA samples were collected from 15 large-scale CFPPs during January to October, 2015.The same sampling method was adopted according to Li et al. (2016).The samples were collected from bottom ashes of electrostatic precipitators (ESP) using a stainless steel spade.The CFPP locations and corresponding administrative divisions where they are located are as shown in Fig. 1.Finally 1 kg of FA was collected from electrostatic precipitators of each CFPP and stored.All samples were dried using a vacuum freeze dryer, and then stored in brown glass bottles before being re-suspended for PM 2.5 and PM 10 fractions.
The particles with aerodynamic diameters less than 2.5 µm (PM 2.5 ) and 10 µm (PM 10 ) were re-collected from fly ashes using re-suspension method (Chow et al., 1994;Kong et al., 2011).Fig. 2 showed the illustrative diagram of resuspension chamber.For each pared of PM 2.5 and PM 10 , 0.5 g of FA was blow out from dust feeding bottle using a dust feeding pump and injected into the re-suspension chamber and then through particle size cutters (2.5 and 10 µm) and intercepted by Teflon filter (diameter 47 mm; Pall Co. USA).Suspension time was 5 min for each sample, and then 40 min was needed to eliminate the interference of last suspension sample.All the Teflon filters were equilibrated for 72 h in a room with constant temperature and humidity before and after sampling, and then weighed.The sampling amount was calculated based on the mass difference of the filters before and after sampling.The sampling amounts ranged from 1110 to 8030 µg for PM 10 and 1150 to 6040 µg for PM 2.5 , respectively.

Particle Size Analysis for FAs and Element Analysis within PM 2.5 and PM 10 Fractions
The particle size distribution of FA was characterized by laser-based analyzer (BT-9300H, Suzhou Qile Electronic Technology Co., China) and water was used as medium.The 30 IEs including Li, Be, Na, P, K, Sc, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Y, Mo, Cd, Sn, Sb, Cs, La, Ce, Sm, W, Tl, Pb, Bi and Th were analyzed by ICP-MS system (Agilent 7500a, Agilent Co. USA).The rest 9 IEs such as Si, Al, Ca, Mg, Fe, Ti, Ba, Sr and Zr were analyzed by ICP-OES system (Agilent 5100, Agilent Co. USA).
For ICP-MS analysis, the 1/2 of Teflon filter was cut into pieces by a ceramic scissors and stored in a Teflon crucible.The 5 mL of aqua regia (V:V = 1:5) and 1 drop of HF acid (V:V = 1:1) were added and heated at 120°C for 2 h and then raised to 130°C until it evaporated to dryness.Then 10 mL 2% of HCl acid was added and heated for 20 min and the extract was transferred into a plastic comparison tubes before analysis.For ICP-OES analysis, the remaining 1/2 filter was cut into pieces and stored in a Teflon crucible, then heated at 300°C for 40 min in a muffle furnace, then gradually heated up to 530-550°C for ashing completely.A few drops of absolute ethanol and 0.1-0.2g NaOH were added and heated for 10 min at 500°C in a muffle furnace.Water was added and boiled on an electric heating plate and  then transferred into a polyvinyl chloride colorimetric tube with 2 mL HCl (V:V = 1:1) added.The extract was finally diluted to 10 mL with high pure water before analysis.

Size Distribution of CFAs from 15 CFPPs
Particle size is one of the most critical features for fly ashes from CFPPs (Li et al., 2014).Fine sized particles, especially PM 2.5 and PM 10 can easily enter into the atmosphere by wind power during the storage of FAs and subsequently result in human respiratory system diseases (Harris and Davidson, 2009;Kong et al., 2011;Zhang et al., 2012;Kulshrestha et al., 2014;Li et al., 2014;Zhang et al., 2015).Table 1S and Fig. 3 list and show the particle size distribution of FAs from 15 CFPPs.In ordinate of Fig. 3, the volume (%) was the volume percentage of particles with specific size.Results indicated that 50% particles for 15 CFPPs ranged from 12.2 µm for FA15 to 55.1 µm for FA6, 75% particles were of 31.5-95.2µm and 90% particles were of 57.5-142 µm indicating a bimodal distribution.As shown in Fig. 3, there were two peaks in each particle size distribution curve.Generally these curves were firstly peaked at about 0.76-1.18µm and secondly peaked at about 21.1-84.9µm for 15 FA samples, the particle size span for second one was wider compared to the first one.The same bimodal distribution pattern was also reported elsewhere (Lind et al., 2003;Arkar et al., 2012;Li et al., 2014; Verma   et al., 2015).It is generally believed that evaporation, condensation and nucleation of inorganic components in feed coal is responsible for the first peak, while the second peak was results from fusion and coalescence of inorganic components (Linak and Wendt, 1994;Yu, 2004;Li et al., 2014).Another hypothesis was proposed by Lin et al. (2002), whereby they asserted that the injection of limestone into furnace for desulphurization resulted in this bimodal particle size distribution.In addition, Li et al. (2015) reported there were four peaks in size distribution curve for FAs emitted from industrial solid waste incineration.The difference of combustion conditions and feed fuels could be the major reason for these observations.

Contents of 39 Elements and Individual Element for 15 CFPPs
As shown in Fig. 4, the highest Ʃ 39 IEs for PM 2.5 occurred for FA9 (0.56 g g -1 ) and PM 10 occurred for FA1 (0.40 g g -1 ).The Ʃ 39 IEs varied from 0.27 to 0.40 g g -1 for PM 10 and 0.29 to 0.56 g g -1 for PM 2.5 , with the mean value being 0.29 ± 0.05 and 0.40 ± 0.08 g g -1 , indicated the IEs were more enriched in lower sized particles.The same conclusions were drawn by other studies (Xu et al., 1997;Wang et al., 1998;Raclavska et al., 2009;Kong et al., 2011;Li et al., 2014;Liu et al., 2015).Xu et al. (1997) reported the contents for 22 IEs including Al, Fe, Ti, As, Co, Sc, Cu, Ga, Ge, Hf, Mo, Ni, Pb, Th, Sr, Li, U, V, Mo, Y, Zn, Be and Se increased with the decreasing of particle size of FAs from a Nanjing Power plant.Wang et al. (1998) studied the 16 IEs in coal fly ashes from different type of boilers and indicated the most elements increased with the decrease in particle size.Raclavska et al. (2009) indicated the contents for Ti, Fe, Na, Mn, Ca, Mg and P in FAs were in line with this trend.Kong et al. (2011) reported the sum of Zn, Cr, Cu, Pb, Ni, As and Cd in re-suspended dust from building surface followed the order as PM 2.5 > PM 10 > PM 100 .Li et al. (2014) reported Be, Cu, Ni, V, Se, Mo and Cd in FAs from a combined heating and power plant in Anhui also exhibited a similar trend.Liu et al. (2015) reported As, Cu, Pb, Zn and Cd were in accordance with this trend, while Hg did not adhere to this trend due to its high volatility.
The burning up of feed coal and excessive oxygen were more important than the other factors such as coal type and plants design during the formation organic pollutants (Revuelta et al., 1999;Cheruiyot et al., 2015).For IEs, the incomplete combustion of coal may be a key factor.The TOC values were further detected for 15 FAs to assess the relationship between IEs and TOC.Ʃ 39 IEs for PM 10 were well negative correlated with TOC of FAs (R = -0.98,p < 0.005), and were weak correlated with TOC for PM 2.5 (R = -0.61,p < 0.01), indicated the less evaporation amounts of IEs during incomplete combustion of coal.
Fig. 5 shows the average individual element content in PM 10 and PM 2.5 , generally the content distributions of 39 IEs were similar between PM 10 and PM 2.5 .For PM 10 , the IEs contents decreased as shown in the following trend Al The average contents for 39 IEs in PM 2.5 were correlated well with those in PM 10 (R 2 = 0.99, p < 0.005).Higher Ca contents in PM 2.5 and PM 10 in this study indicated the influence of desulphurization of coal by limestone injection into furnace in 15 CFPPs (Li et al., 2014).Li et al. (2014) found CaO content accounts for as high as 6.97% of major oxides due to the addition of limestone, while it was only 1.23% for Chinese coal (Dai et al., 2012).
Among 39 IEs, 13 heavy metals (HMs) such as Pb, Cr, Cd, As, Al, Co, V, Sb, Mn, Sn, Tl, Cu and Zn received more attention due to their adverse health effects to people.
In order to investigate the particle size distribution of 39 IEs, the content ratio of PM 2.5 to PM 10 for each element was calculated and showed in Fig. 6.The PM 2.5 /PM 10 ranged from 0.89 for Cu to 1.72 for Si with the mean value as 1.29 ± 0.13.Except for Cu, PM 2.5 /PM 10 ratios for the remaining 38 IEs were higher than unity, indicating that they were more inclined to be enriched in PM 2.5 compared to PM 10 .The differences in volatilization and adsorption mechanisms among IEs could be the reasons for the observed variations.Seames (2003) reported that evaporated IEs were more easily condensed in particles with high specific surface area.The fine-sized particles had larger surface area, higher adsorption activity and ability compared to the coarse ones and resulted in higher contents of IEs in the fine particles (Xu et al., 1997;Wang et al., 1998;Liu et al, 2015).IEs with relatively higher volatilization, evaporated from coal at high temperature and physically adsorbed by surface of FAs during cooling of flue gas, which was named as volatilization and condensation mechanism (Chen et al., 2001).For elements with lower volatilization, chemical reaction adsorption between element and surface chemicals of FAs would be the main process (Li et al., 2014).

Geo-accumulation Index Assessment for Each Element within PM 10 and PM 2.5
Geo-accumulation index (Igeo) proposed by Muller has been widely used in element studies (Muller, 1969;Wei et al., 2009;Kong et al., 2011;Li et al., 2013).where, C n is the measured content of element in FAs, B n is the background value of elements in soil.The constant of 1.5 allowed evaluating natural fluctuation of a given element in environment and analyzing the slight anthropogenic effects (Wei et al., 2009).In this study, the background values of IEs in soils for different Chinese administrative divisions where CFPPs located were presented in CNEMC.(1990) and listed in Table 2S.As shown in Table 1, the pollution levels of IEs were further identified as seven categories based on their Igeo values.
Among 39 IEs involved in this study, only 37 background values for corresponding IEs were found in CNEMC.(1990), so Igeo values were calculated for these 37 IEs.Generally Igeo values for 36 IEs except for Cu in PM 2.5 were higher than those in PM 10 (Fig. 7).In case of PM 10 , Class B including Tl, U, W, Y, Zn, Be, Li and Sr indicated uncontaminated to moderately contaminated, C including Sb, Pb, Bi and Mo indicated moderately contaminated, D including Cd, Sn and Cu indicated moderately to heavily contaminated state, E including Cr indicated heavily contaminated, G including K, Na, Ti, Fe, Mg, Ca and Al indicated extremely contaminated, and the rest 14 elements belong to A indicated uncontaminated state.For 37 elements in PM 2.5 , 10 IEs as Th, Tl, W, Sm, Ce, La, Y, Ni, Be and Li were belong to B, 6 IEs as U, Zn, Sb, Pb, Sr and Mo were belong to C, 3 IEs as Cu, Sn and Cd were belong to D, Cr was belong to E, 7 IEs as K, Na, Ti, Fe, Mg, Ca and Al were belong to G, and rest 11 IEs were belong to A. For 13 HMs, 8 metals as Tl, Zn, Sb, Pb, Cu, Cd, Cr and Al in PM 2.5 and PM 10 with Igeo > 0 indicated different pollution levels of them.

Similarity Comparison of Element Profiles for PM 2.5 and PM 10 among 15 CFPPs
The fine-sized PM, especially for PM 2.5 and PM 10 , would result in a series of adverse impacts on atmospheric quality and human health, such as visibility reduction, deterioration of ecosystem through deposit of toxic pollutants within them, change of earth radiation budget and global climate and so on (Kchih et al., 2015).PM has become a severe environmental challenge to the present China.As a receptor model, chemical mass balance (CMB) model has been widely used for source apportionment of PM (Moreno et al., 2009;Kong et al., 2010).Accurate source profiles were the basis of the calculation result of the CMB (Kong et al., 2010).
FAs emitted from CFPPs were an important PM source and establishment of element profiles for them from every CFPP was a heavy work.The similarity of element profiles among different CFPPs made they can replace each other, and then simplified this work.So the similarity of element profiles among different PM 2.5 or PM 10 samples was identified by a parameter of coefficient of divergence (CD) in this study.CD was calculated as followed: where j and k were the different sources (refer to different CFPPs in this study), p was the number of analyzed elements, and x ij and x ik were the mean masses of component i (refer to 39 elements in this study) for j and k (Wongphatarakul  et al., 1998;Li et al., 2016).The CD value of 0 indicated source profiles of j and k were completely same, 1 indicated that they were completely different, and less than 0.3 suggested they were similar (Wongphatarakul et al., 1998).
In this study, CD values for 105 pared of element profiles for PM 2.5 or 105 pared of those for PM 10 were calculated to investigate the similarity of PM 2.5 or PM 10 between any two CFPPs.Results showed that there were only 8 points higher than 0.30 among 105 points for PM 2.5 , and were only 10 points higher than 0.30 among 105 points for PM 10 (Fig. 8).The CD values for 105 pared of data were 0.254 ± 0.038 for PM 2.5 , and 0.244 ± 0.054 for PM 10 , indicated the similarity of element profiles of PM 2.5 or PM 10 from 15 CFPPs.In short, the element profiles for PM 2.5 or PM 10 may be substituted for each other.This discovery would significantly simplify the work of element profiles of re-suspended particles from ash ponds.

Enrichment Factors Analysis
To evaluate the behavior of 39 IEs during coal burning process, a parameter named as relative enrichment factor (REF) was proposed by Meij (1994) and calculated for each element as follows: where A refer to element content in FAs (mg kg -1 ), B refer to that in coal (mg kg -1 ), and C refer to ash content of coal (%).REF values were in the range of 0.70-1.30indicated no significant enrichment and exhaustion occurred, higher than 1.30 indicated the significant enrichment occurred, and lower than 0.70 indicated the exhaustion occurred (Meij, 1994;Lu et al., 2009).Li et al. (2014) reported there were three categories according to REF values: 1) elements with REF around 1 indicated they were not evaporated during burning process; 2) with REF around 0.7 indicated they were evaporated during burning process; 3) with a very low REF indicated they were completely evaporated during burning process.Dai et al. (2012) reported the background contents of 51 elements and 10 element oxides for Chinese coal.The average ash content of 21.5% for Chinese coal was also reported.Among these elements, 28 elements were involved in this study such as Li, Sc, V, Cr, Co, Ni, Cu, Zn, As, Rb, Sr, Y, Zr, Mo, Cd, Sn, Sb, Cs, Ba, La, Ce, Sm, Tl, Pb, Bi, Th, U and W with their contents (reported in mg kg -1 ) as 31.8, 4.38, 35.1, 15.4, 7.08, 13.7, 17.5, 41.4, 3.79, 9.25, 140, 18.2, 89.5, 3.08, 0.25, 2.11, 0.84, 1.13, 159, 22.5, 46.7, 4.07, 0.47, 15.1, 0.79, 5.84, 2.43 and 1  CFPPs were analyzed using ICP-MS and ICP-OES.The sum of 39 IEs within PM 2.5 (0.292-0.564 g g -1 ) were higher than those of PM 10 (0.269-0.403 g g -1 ), indicated the IEs were more inclined to be enriched in fine sized particles.
For individual element content, all the 39 IEs except for Cu and Sb, were higher in PM 2.5 .The content ratios in PM 2.5 to PM 10 for 39 IEs except for Cu were higher than 1.0, and ranged from 1.10 for Sb to 1.72 for Si.Higher levels of Cr for all the PM 2.5 and PM 10 should be paid more attentions.The analysis result for particle size distribution of FAs from 15 CFPPs indicated 50% particles for 15 CFPPs ranged from 12.2 µm for FA15 to 55.1 µm for FA6, and 90% particles were of 57.5-142 µm.All the 15 size distribution curves had two peaks with the first one occurred at 0.762-1.18µm and second one occurred at 21.1-84.9µm.
Geo-accumulation index values for 23 IEs in PM 10 and 26 IEs in PM 2.5 were higher than threshold value of 0, indicated their pollution levels ranged from moderately to extremely contamination for them.Igeo values for 8 heavy metals including Tl, Zn, Sb, Pb, Cu, Cd, Cr and Al in PM 2.5 and PM 10 were higher than 0, especially for Cr, it was 3.57 and 3.61 for PM 10 and PM 2.5 .
The element profiles of PM 2.5 between any two CFPPs may be substituted for each other based on lower coefficient of divergence as 0.254 ± 0.038.The same situation was occurred for PM 10 (0.244 ± 0.056).The relative enrichment factors (REF) for 35 and 33 elements in PM 10 and PM 2.5 were less than 0.70 or in the range of 0.70-1.30,indicated evaporation, no condensation or weak condensation processes occurred for them during the coal combustion.4 IEs including Cr, Cu, Zn and Sn in PM 10 , and 6 IEs including Cr, Cu, Zn, Sn, W and Pb in PM 2.5 had the REF values higher than 1.30 indicated significant condensation occurred for them.Cr had the highest REF as 17.4 and 18.3 for PM 10 and PM 2.5 , respectively.
Fig. 8. Box chart for coefficient of divergence values of PM 2.5 and PM 10 .

Table 1 .
Seven categories of Igeo value for each trace element.Igeo values for each element in PM 10 and PM 2.5 .

Table 2 .
Relative enrichment factors of 39 elements in PM 10 and PM 2.5 .Tl, Pb, Bi, Th, Si, Al, Ca, Mg, Fe, Ti, Ba, Sr, Zr and U within PM 2.5 and PM 10 fractions of FAs collected from 15