Size Distribution of Inorganic Elements in Bottom Ashes from Seven Types of BioFuels across Beijing-Tianjin-Hebei Region , China

A total of fifty-six bottom ash (BA) samples from the indoor burning of seven bio-fuels (BFs) using a zaotai stove were collected from eight sampling sites across the Beijing-TianjinHebei (BTH) region from April till December in 2016. Each one was divided into six parts as PM+148, PM93–148, PM67–93, PM53–67, PM40–53 and PM–40 using a vibrating screen. The three parts containing the smallest particles, consisting of PM53–67, PM40–53 and PM–40, were selected for analysis of 39 inorganic elements (IEs) using ICP-MS and ICP-OES. The firewood–walnut (WAL) sample had the lowest ash yield as 34.3 ± 3.55 g kg; the corresponding values for the 6 crop straws were millet (MIL) > sorghum (SOR) > sesame (SES) > corn (COR) > cotton (COT) > soybean (SOY). The ash yields, in general, positively correlated with the particle sizes of the BAs for all seven BFs. The top eight elements (TEs), namely Si, Ca, Mg, K, P, Al, Na and Ti, dominated in all BAs regardless of their fluctuation among the BFs, and Ʃ8TEs were well correlated with Ʃ39IEs (R = 0.99, p < 0.001). The Ʃ39IEs were not correlated with the particle size of the BA due to the significant fluctuation of these TEs. The trace elements, namely Sc, Li, P, V, Cr, Co, Ni, Sb, As, Y, Cs, Bi, Tl, Th, Sn, Cd, La, Ce, Sm, W and U, were negatively well correlated with the particle sizes of the BAs, while the TEs, except for Al, did not display this trend. The heavy metal (HM) profiles were similar between any two sizes of BA for all BFs based on lower values of the coefficient of divergence (CD); with higher CD values, 6 out of 21 pairs of BFs had different HM profiles. All 7 BFs produced higher levels of Zn, Cu, Pb, V, Cr and Ni than other HMs. The content of V in BAs from COR was beyond the limit designated by European countries for BAs used as soil conditioner. The enrichment factors (EFs) of 12 HMs were not correlated with the particle sizes of the BAs, due to the fluctuation of Al, which was used as a reference element among differently sized BAs. The HMs Cu, Zn, Cd and Sn had higher EFs (more than 10), indicating that they were significantly influenced by human activities.


INTRODUCTION
Biomass burning (BB) was widely known as open burning of any non-fossilized agricultural residues, grass and forest, and residential combustion of BFs for cooking and heating (Engling et al., 2014).It is the largest source of fine carbonaceous particles and the second largest trace gases such as SO 2 , NO x , NMVOCs, NH 3 , could affect on atmospheric chemistry, global climate and human health (Arola et al., 2007;Liu et al., 2013;Popovicheva et al., 2015;Tian et al., 2015;Lee et al., 2016;Zhu et al., 2016;Li et al., 2017a;Ommi et al., 2017).Bond et al. (2004) also reported BB was the largest source of primary organic carbon (POC) and element carbon (EC) in the world.The serious atmospheric pollution attributed to various pollutants from BB were reported frequently (Eckhardt et al., 2007;Alves et al., 2010;Kondo et al., 2011;Engling et al., 2014).Some studies revealed fine particles can cause adverse health effects such as pulmonary function failure, respiratory diseases and lung cancer (Huang et al., 2012;Shi and Yamaguchi, 2014;Zhou et al., 2017).The toxic pollutants adsorbed and absorbed in emitted particles from BB such as heavy metals and organic compounds could result in seriously adverse health effects (Habeebullah, 2016;Wei et al., 2016;Li et al., 2017b).
The people relying on the traditional use of biomass for cooking and heating was predicted to rise from 2.7 billion today to 2.8 billion in 2030 around the world (UNDP, 2002).
In China, the annual straw burning capacity was as high as 140 Tg and accounted for 23.3% of total crop products, especially in eastern and northeastern China (Cao et al., 2008).BB as renewable energy was commonly used in Chinese rural areas and domestic BB was the most prevailing utility in China (Wang et al., 2009;Zhou et al., 2017).During combustion process, the IEs stored in biomass and soil for a long time emitted again into atmosphere and subsequently transported to other areas (Eckhardt et al., 2007).The re-emission of radioactive cesium-137 and mercury from biomass burning were documented (Sigler et al., 2003;Wotawa et al., 2006).The much more BAs compared with fly ashes produced from BB were not properly treated and always casually dumped in China.It was a fact can't be ignored the pollutants in dumped BAs could enter into soil and atmosphere by wind strength.But few comprehensive studies on emitted element characteristics in BAs from indoor BB were conducted in China.
As a cultural and political center of China and economic center of northern China, BTH suffered extreme and frequent haze episodes caused by rapid economic growth and urbanization (Zhang et al., 2017).Two cities in Hebei province as Baoding and Shijiazhuang were ranked as 1 st and 2 nd most polluted cities in China (Balasubramanian et al., 2017).Annual PM 2.5 level of BTH in 2014 was 6.2 and 2.7 times as much as the corresponding values appointed by Chinese Class I and Class II standards (Chen et al., 2017).Some studies reported the BB contributed 35-50% of fine particles to Beijing atmosphere.Consequently, it is of great importance to clarify the pollution characteristic of heavy metals in BAs from BB.
Zaotai stove is a typical burning tool for indoor BB in BTH.In this study, 8 sites across BTH were selected to collect BAs for 7 BFs from Zaotai stove and the main aims were to investigate: 1) influence of BF types on elemental compositions in BAs from BTH; 2) size distribution of IEs for 7 BFs; 3) similarity comparison of heavy metal profiles among different BFs or BAs with different sizes for each BF.

Sampling Area Description and Bottom Ash Sampled
Beijing-Tianjin-Hebei (BTH) region (36.07°N-42.65°Nand 113.46°E-119.79°E)locates in North China plain, which includes 2 municipalities (Beijing and Tianjin) and 11 prefecture-level cities of Hebei province, covers an area of 216,500 km 2 and has 110,000,000 people.BTH is the most important and dynamic economic zone in northern China, but also is the most air polluted zone (Chen et al., 2017;Zhang et al., 2017).Although BAs were indirectly entered into air, but it could entered into air by wind power after dumped carelessly.To our knowledge, few studies were conducted on BB across BTH.
In this study, 8 sites including Beijing, Tianjin and other 6 prefecture-level cities such as Tangshan, Xingtai, Zhangjiakou, Chengde, Baoding and Shijiazhuang were selected to collected BA samples for 7 BFs from Zaotai stoves (Fig. 1).The 56 BAs were vibration screened using vibration sieve (Xinxiang Beiteli Vibration Machine LTD.) into 6 different sized parts.Finally the lower sized 3 parts including PM 53-67 , PM 40-53 and PM -40 were selected for 7 BFs for analysis of 39 IEs because the finer parts were more easily enter into atmosphere compared with the coarse parts.All crop straw and firewood samples were dried using a vacuum freeze dryer, and then stored in brown glass bottles before analysis in order to eliminate the influence of water.

Sample Pre-treatment and Element Analysis
The elements in BAs were analyzed using combination of ICP-MS system (Agilent 7700X, Agilent Co. USA) and ICP-OES system (VISTA-MPX, Varian Co. USA).The 30 of 39 IEs such as 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 and the other 9 ones such as Si, Al, Ca, Mg, Fe, Ti, Ba, Sr and Zr were analyzed by ICP-OES.The same analysis method was selected in this study as Li et al. (2017b).
The method was described briefly as followed: 1) For ICP-MS, 0.2000 g BA was heated together with 5 mL of aqua regia and 1 drop of HF acid until evaporated to dryness.Then was heated for 20 min with 2% of HCl acid and the extract was transferred into a plastic comparison tubes waiting for analysis.2) For ICP-OES, 0.1000 g of BA was transferred into a Teflon crucible and heated using a muffle furnace, and then gradually heated up to 530-550°C for completely ashing.The absolute ethanol and NaOH powder were added and heating again at 500°C for 10 min in a muffle furnace.It was subsequently boiled together with added water on an electric heating plate.The HCl acid was added and transferred into a PVC colorimetric tube and diluted to 10 mL waiting analysis.
Standard soil materials (SSMs) as GBW07446-GBW07457 (Center for National Standard Matter, China) were used to assess the accuracy of ICP-MS and ICP-OES.The background contamination was routinely assessed using process and solvent blank experiments.The standard materials were analyzed every 5 BA samples to evaluate the instrument repeatability.One BA sample was repeated analyzed every 6 samples with the RSD values were lower than 15%.

Size Distribution of Ash Yields for Seven BFs
The BA samples were vibration screened into 6 parts and weighed to obtain the corresponding ash yield values.Fig. 2(a) showed the size distribution of ash yields for 7 BFs.In generally, ash yields for all the 7 BFs showed the positive correlation with particle sizes.PM +148 accounted for the highest proportion of total ash amounts and PM -40 had the lowest corresponding values for all the 7 BFs.The percentage of PM +148 in total ash amounts were ranged from 33.53 ± 4.12% for WAL to 66.02 ± 5.26% for SES, while the corresponding values for PM -40 were in the range of 0.84 ± 0.112% for SOY to 5.89 ± 0.652% for COT.The sum of ashes for 3 lower sized parts including PM 53-67 , PM 40-53 and PM -40 were ranged from 7.91 ± 0.521% (WAL) to 22.05 ± 3.123% (COR) of the total ash yields.

Contents of Individual Elements in BAs
Total contents of 39 inorganic elements (Σ 39 IEs) (in mg g -1 ) were ranged from 0.249 ± 0.021 (PM 40-53 from WAL) to 0.493 ± 0.036 (PM 53-67 from SES) with the mean value as 0.412 ± 0.071 for 7 BFs (Fig. 3(a)).The top elements (TEs) including Si, Ca, Mg, K, P, Al, Na and Ti dominated in BB samples for all 7 BFs regardless of their contents varied significantly among different BFs (Table 1(a)).The Σ 39 IEs were not correlated with particle size of BAs for 7 BFs due to obvious fluctuation of 8 TEs among 7 BFs.The sum of these 8 TEs (Σ 8 TEs) (mg g -1 ) for all sized BAs from 7 BFs were ranged from 0.247 ± 0.021 (PM 40-53 from WAL) to 0.491 ± 0.032 (PM 53-67 from SES).The Σ 8 TEs accounted for 98.76% of Σ 39 IEs for PM 53-67 from WAL to 99.48% of Σ 39 IEs for PM 53-67 from SOR. Fig. 2(b) showed that Σ 8 TEs for all BAs were well correlated with the corresponding Σ 39 IEs (R 2 = 0.99, p < 0.001) (Fig. 3(b)).These TEs varied significantly not only among different BFs, but also among different sized BAs for the same BF.For PM 53-67 among 7 BFs, the contents for 8 TEs followed as K > Ca > P > Si > Mg > Fe > Al for COT, while they were Si > K > Ca > Mg > P > Fe > Al, Si > Ca > K > Fe > P > Al > Mg, Si > Ca > K > Mg > Fe > P > Al, Ca > Si > K > P > Al > Mg > Fe, K > Ca > Al > Mg > Si > P > Fe, and K > Ca > Si > Mg > P > Al > Fe for COR, MIL, WAL, SOR, SOY and SES, respectively.The contents of Na were higher than those of Ti for all the 7 BFs (Table 1(a husk combustion, while Pb, Fe and Mg dominated in BAs from a mixed BF (straw, bamboo, cow dung, leaves and plants).Wang et al. (2016) reported the main elements were also K, Ca, P and Mg in BAs from burning of straws of corn, wheat and rice due to they were essential ones for plant growth.
K was often used as identification element for biomass burning (Duan et al., 2004).In this study, the content of K varied significantly among 7 BF from different sampling sites, possibly resulted from the difference of plant physiological characteristics and local soil characteristics among different sites in BTH (Duan et al., 2004;Wang et al., 2016).Consequently the use of K identification element should be cautioned in BTH.
It should be noted, the IEs were classified into 3 classes according to their relationship with particle sizes: 1) IEs with higher contents (e.g., Si, Ca, K, Mg, Fe, P and Ti) except for Al were not correlated with particle size for all the 7 BFs, Al was well correlated with particle size of BAs from 7 BFs; 2) IEs with lower contents (e.g., Sc, Li, P, V, Cr, Co, Ni, Sb, As, Y, Cs, Bi, Tl, Th, Sn, Cd, La, Ce, Sm, W and U) were well negative correlated with particle size for all the 7 BFs; 3) IEs with medium content (e.g., Mn, Na, Zn, Cu, Sr, Zr, Rb, Pb and Mo) within partial BFs showed the well correlation relationship (Tables 1(a)-1(c)).For examples, Pb within 5 of 7 BFs were well correlated with particle size of BAs such as COT, MIL, WAL, SOY and SES, while Mn within only 2 of 7 BFs showed this trend such as COR and MIL.Li et al. (2017b) reported the 39 IEs except for Cu within fly ashes emitted from 15 Chinese power plants were more inclined to enrich in finer particles as PM 2.5 .These differences possibly resulted from the different ash characteristics, combustion conditions, fuels and combustion devices.

Suitability Assessment of Application of Bottom Ashes as Soil Conditioner
The exposure of people to heavy metals (HMs) would pose serious adverse health effects (Zheng et al., 2016).Cr and Mn abounded in particles could result in acute and chronic damage of lung (Zheng et al., 2013).Cr and Cd were strongly carcinogenic elements regardless of their low contents (Zheng et al., 2016).
In this study, 12 HMs (e.g., Fe, Mn, Cu, Zn, Cr, Cd, Sn, Co, As, Pb, V and Ni) most concerned by people were all detected.The HMs including Mn and Fe dominated among all the BAs.The HMs including Zn, Cu, Pb, V, Cr and Ni had higher levels among all BAs (Tables 1(a) and 1(c)).Saqib and Bäckström. (2016) also reported Zn, Cu and Pb were predominant HMs in fly ashes from combustion of 13 BFs, while Lanzerstorfer, (2015) reported the top HMs in fly ashes from 8 Australian grate-fired BB plants were Zn, Pb and Cr.Although the content order was similar, the contents of corresponding HMs were much lower than those of fly ashes from biomass burning.The volatile HMs would rather enrich in finer fly ashes compared to coarser BAs possibly be the explanation (Lanzerstorfer, 2015).
In Australia, BAs from BB were widely used as soil conditioner if the contents of contained HMs were below the corresponding threshold values, while fly ashes were prohibited for that utility way due to higher contents of contained HMs (Lanzerstorfer, 2015).Emilsson (2006) and Stupak et al. (2008) reported the limits (in mg kg -1 ) of HMs such as As, Cr, Cd, Cu, Hg, Ni, Pb, V and Zn designated by two European countries, they were 30,30,100,400,3,70,300,70 and 7000 for Sweden,and 30,30,100,400,3,70,300,70 and 7000 for Lithuania,respectively.In this study, the contents (in mg kg -1 ) of Pb ranged from 4.27 ± 0.65 (PM 53-67 from SOY) to 28.3 ± 5.26 (PM 53-67 from COR), and V ranged from 6.03 ± 1.02 (PM 53-67 from WAL) to 81.1 ± 9.25 (PM -40 from COR).Cd had the lowest contents ranged from 61.5 ± 6.22 ng g -1 (PM 53-67 from WAL) to 610 ± 85.6 ng g -1 (PM -40 from COR) (Table 1(c)).Among 12 HMs in all sized BAs from 7 BFs, only V in COR exceeded the limit of 70, so the application of COR BAs in BTH as soil conditioner should be prohibited.

Similarity Comparison of Heavy Metal Profiles
As a receptor model, chemical mass balance (CMB) model was often used to the work of source apportionment of atmospheric pollutants for different regions (Kong et al., 2011).The source profiles for different emission sources were urgently needed for CMB calculation, while the establishment of source profiles was a heavy work (Li et al., 2017b).The similarity of element profiles among different BFs or different sized BAs within one BF would make they can be replaced each other and consequently reduced workload of establishment of source profiles (Li et al., 2017b).
A parameter as coefficient of divergence (CD) was commonly used to compare the similarity of pollutant profiles emitted from different sources (Wongphatarakul et al., 1998;Kong et al., 2011;Li et al., 2016Li et al., , 2017b)).A threshold of CD as 0.3 was often used to identify the similarity of pollutant profiles of source j and k, less than 0.3 indicated pollutant profiles between j and k were similar, higher than 0.3 suggested they were different, while 0 and 1 indicated they were completely same and different, respectively (Wongphatarakul et al., 1998).Kong et al. (2011) reported the PAH profiles for six stationary sources in a Chinese oilfield city-Dongying were similar based on lower CD values (lessen than 0.3).
Source apportionment of atmospheric particle bounded HMs was more concerned due to their serious threat to human health.The CD values for HM profiles among any 2 out of 7 BFs and among any 2 different sized BAs within one BF were calculated as followed: where p was the number of analyzed HMs and it was 12 in this study, j and k were emission sources of HMs (they were BA samples from 7 BFs or different sized BAs from the same BF in this study), x ij and x ik were mass percentage of element i within j and k (Wongphatarakul et al., 1998;Li et al., 2016Li et al., , 2017b)).
The CD values between any 2 sized BA samples within one BF were listed in Table 2.All the CD values were lessen than 0.3 and they were ranged from 0.025 ± 0.015 (PM 53-67 vs PM 40-53 within COT) to 0.259 ± 0.012 (PM 53-67 vs. PM -40 within WAL) with the mean value as 0.144 ± 0.068, indicated HM profiles between any 2 sized BAs within   2).Consequently the HM profiles for different sized BAs within one BF were similar and could be replaced each other.Due to the similarity of HM profiles between any 2 sized BAs for one BF, HM profile of PM 53-67 was appointed to represent the corresponding profile for whole BF.CD values were calculated based on 12 HMs for any 2 BFs and the results were listed in Table 3.In generally, the CD values between any 2 BFs were higher than those between any 2 sized BAs within one BF.Except for 6 bold numbers, the other 15 CD values were all lessen than 0.3 (Table 3).The highest CD value of 0.467 ± 0.067 occurred at COR vs SOY and the lowest one as 0.153 ± 0.023 occurred at MIL vs. SOR.The bold numbers higher than 0.3 suggested the source profiles for these sources were different each other.The HM profile of COT were similar with those of the other 6 BFs based on 6 CD values lessen than 0.3.WAL was similar to HM profiles of 5 BFs except for SES.SOR was similar to those of 5 BFs except for SES.COR was similar to those of 4 BFs except for SOY and SES.MIL was similar to those of 4 BFs except for SOY and SES.

Enrichment Factors of Heavy Metals in all BAs
The enrichment factor (EF) was widely used to identify the element was originated from human activities or natural process, and then assess the content of anthropogenic influence.The element was mainly crustal origin if it's EF close to 1, while EF higher than 10 suggested significant influence of human activities (Li et al., 2013).In this study, EF was used to assess the influence of casually dumped BAs on its receptor soil or air.The EF was calculated based on following equation: where C n is measured content of element in BA, B n is background content of element in soil from BTH, C ref and B ref is measured content and background content of reference element in BA and BTH soil.Although the utility of different reference elements would result in the different EF values, Al was often used as reference element in Chinese related studies (Hao et al., 2009;Zhang et al., 2010).So Al was chosen as reference element accordingly in this study.Fig. 4 showed the EFs for 12 HMs in different sizes BAs from 7 BFs.The EFs for 12 HMs varied significantly among different BFs and were not correlated with particle size of BAs for all the 7 BFs.The EFs for Cu and Zn were higher than 10, indicated the significant influence of dumped BAs on receptor soil.The highest EF as 69.1 for Cu occurred at PM 53-67 from SOY.
For 12 HMs, Cu, Zn, Cd, Sn and Pb had higher EFs among all the 7 BFs (Fig. 4).The EFs for HMs varied significantly among different BFs.For examples, Cu was more inclined to enrich in SOY, while As and Pb were more enriched in COT.All the EFs for 12 HMs among 7 BFs showed not correlated with particle size of BAs regard less of their contents had well correlation with particle sizes, which was resulted from varied Al contents among different sized BAs.

CONCLUSIONS
BA samples were collected for 7 BFs from 8 sampling sites across the BTH region, resulting in a total of 56 samples, each of which was divided into 6 parts based on particle size.The 3 parts containing the smallest sizes, PM 53-67 , PM 40-53 and PM -40 , respectively, were collected for analysis of 39 IEs using ICP-MS and ICP-OES.The Σ 39 IEs (reported in mg g -1 ) ranged from 0.249 ± 0.021 to 0.493 ± 0.036 with the mean value as 0.412 ± 0.071 for these BFs.The IEs K, Ca, Si, P, Fe, Al, Na and Ti dominated in the BAs despite their content varying significantly among the 7 BFs.The predominant IEs, except for Al, did not correlate with the particle size of the BA for any BF, while the trace IEs-Sc, Li, P, V, Cr, Co, Ni, Sb, As, Y, Cs, Bi, Tl, Th, Sn, Cd, La, Ce, Sm, W and U-negatively correlated with the particle size of the BA for all BFs.All the BAs, except for V in COR, had less HM content than the limit designated by European standards for soil conditioner; hence, BAs from COR should be prohibited in the BTH.
Comparing the content of various HMs, all of the BAs possessed higher levels of Zn, Cu, Pb, V, Cr and Ni.For 12 HMs, CD values below 0.3 for any two sizes of BA from the same BF indicated the similarity in HM profiles between differently sized BAs from one BF.CD values above 0.3 resulted in different HM profiles for 6 of 21 BF pairs.The HMs Cu, Zn, Cd, Sn and Pb possessed higher EF values among the 7 BFs.The EF values for Cu, Zn, Cd and Sn were higher than 10, suggesting the significant influence of human activities.

Table 1 (a).
Concentration of individual element in BAs with different sizes (µm) from 7 BFs across BTH.Concentration of individual element in BAs with different sizes (µm) from 7 BFs across BTH.

Table 1 (
c). Concentration of individual element in BAs with different sizes (µm) from 7 BFs across BTH.

Table 2 .
CD values for HM profiles between any two sized BA samples within the same BF.

Table 3 .
CD values for HM profiles among any two out of seven biofuels