Distribution and Sources of Polychlorinated Biphenyls in Air, Dust, and Sediment from India

Persistent organic pollutants, such as polychlorinated biphenyls (PCBs), pose a serious risk for human health and the environment. In this study, PCBs contamination and sources of ambient air, road dust, and sediments in the most polluted city in India, Raipur has been measured over the period 2008–2015. The seasonal variations of particulate matter (PM), elemental carbon (EC), organic carbon (OC), and carbonate carbon (CC) were studied, and maximum concentrations were detected in the December–January period each year. Total PCBs concentrations in the ambient air (associated to particulate matter), road dust, and sediments samples during 2008 were in the 186–645 pg m, 102–537, and 241–538 ng g range, respectively. 2-chlorobiphenyl (PCB-1) and 4-chlorobiphenyl (PCB-3) were the dominant chemical compounds identified. A substantial vertical migration of the PCBs in the sediment was observed. Concentration variations (spatial and temporal), correlations, and sources of PCB are discussed. In particular, an average increment rate of 6.2%, 4.9%, and 5.4% of PCBs concentration in the particulate matter (PM10), road dust, and sediments respectively, was observed over the 2008–2015 period. The reported data points to India’s low degree of accomplishment of the Stockholm Convention’s requirement to phase out the use of PCBs in equipment by 2025 and ensure elimination of PCBs by 2028. Author keywords: PCBs; Contamination; Air; Dust & sediment; Toxicity; Sources.


Introduction
Persistent organic pollutants (POPs) (i.e., aldrin, chlordane, DDT, dieldrin, endrin, heptachlor, hexachlorobenzene, mirex, toxaphene, polychlorinated biphenyls (PCBs), polychlorinated dibenzo-p-dioxins (PCDD), and polychlorinated dibenzofurans (PCDF), plus the 16 so-called "new POPs" under the Stockholm Convention, are characterized by their high half-lives, low water solubility, and high lipid solubility (Harrad 2009). In particular, PCBs, a group of man-made oily liquids or solids, are widely used in electrical equipment, hydraulic fluids, heat transfer fluids, coolants, lubricants, and plasticizers in paint, paper, and plastics due to their low electrical conductivity and high resistance to heat and thermal degradation (Robertson and Hansen 2001). PCBs are a class of aromatic compounds having two benzene rings with a maximum of 10 substituted Cl-atoms, and are mixtures of up to 209 individual chlorinated compounds, known as congeners, whose chemistry differs from species to species (Hutzinger 1974). PCBs have been reported to cause many adverse effects both on wildlife and on human health, including immune deficiency, nervous system alteration, endocrine disruption, and gastrointestinal system bleeding and liver damage (ATSDAR 2014).
Contamination with PBCs has been reported in air, water, soil, dust, and sediment samples from all over the world Biterna and Voutsa 2005;Chakraborty et al. 2013Chakraborty et al. , 2016Devi et al. 2014;Goel et al. 2016;Kim and Masunaga 2005;Kumar et al. 2011;Nasir et al. 2014;Sakin and Tasdemir 2016;Syed et al. 2013;Wang et al. 2013). In fact, at present, the United Nations Environment Program (UNEP) estimates that 83% of the total amount of PCBs in the world (ca. 14 million tons) still remains to be eliminated.
In the particular case of India, the Stockholm Convention was ratified in 2002 and entered into force in 2006. Preliminary investigations on PCBs contamination in sediments and sludge from the Raipur area has been previously reported (Patel et al. 2013(Patel et al. , 2015. Findings of the highest concentrations of the monochlorobiphenyl (MCBP) congeners were reported to date. Nonetheless, as noted by the PCB Elimination Network (PEN), "data on elimination of PCB is often incomprehensive, outdated, and incomplete (United Nations Environmental Programme, UNITAR 2017)." Consequently, the aim of present work is to describe the distribution, variation, sources, and fate of PCBs congeners in different environmental samples (air, dust, and sediment) from the most industrialized area of central India, Raipur city (capital of Chhattisgarh state), to strengthen the analyses of India's PCBs situation, in line with the "country specific diagnoses" PEN key theme. 1

Quality Assurance/Quality Control Analysis
One-way analysis of variance (ANOVA) was performed to analyze the distribution variability of the PCBs concentrations in the particulates, road dust, and sediments. IBM SPSS 20.0 (SPSS Inc.) software was employed for the factor analysis of source contributions of PCBs in the air, dust, and sediment by extracting factors with an Eigenvalue of >1.0 (Shyu et al. 2011).
The surrogate standard ( 13 C 12 -labelled PCBs) was used to determine the extraction efficiency for the targeted PCB congeners. The standard was added prior to the extraction processes, and the average recoveries of these standards from the different environmental matrices varied from 91.1% to 102%. The calibration curve was prepared by injections of standard solutions containing a mixture of the PCBs at four concentration levels. The blanks used were clear of all the examined PCBs. The limits of detection (LODs) and quantification (LOQs) were determined. Reported values are expressed as an average across three replicate measurements, both for the carbon and PCB analyses.

Study Area
The capital city of Chhattisgarh state, Raipur (21°23 ′ N, 81°63 ′ E), with a population of ca. 2 million inhabitants, was selected for the proposed investigation due to its high degree of industrialization. Raipur city and the nearby region are home to many coal, power, steel, and aluminum industries. For instance, Asia'sbiggest steel plant is located in Bhilai, and Korba (known as the "power capital of India") is heavily polluted by thermal power plants (Jaiswal et al. 2019a, b).

Environmental Samples Collection
The sampling of the particulate matter was carried out for 13 months (February 2007to January 2008 in Kota, a residential area (Fig. 1). For spatial variation studies, the sampling was carried out at three additional locations: Pt. Ravishankar Shukla University, Amapara, and Raipura in December 2008. These sampling sites were approximately 2 km distant from each other. The locations at Kota and Amapara are located at commercial and traffic sites. The distance of these sampling sites from the industrial area was approximately 2 km to the east. For temporal variation studies, Kota site was selected, and one PM 10 sample was collected every January from 2009 to 2015.
A Partisol model 2300 sequential speciation air sampler (Thermo Scientific, Waltham, MA, USA) was used for the collection of particulate matter (PM) samples. Coarse particles with a diameter <10 μm (PM 10 ) were collected on dried Whatman QM-A 47-mm quartz fiber filters, housed in molded filter cassettes. The sampler was installed on the roofs of buildings at approximately 10 m above the ground level. The sampling was carried out for 24 h (6 a.m.-6 a.m.). Similarly, a relevant field sample blank was prepared, as the filter paper was exposed to the environment during the mounting and dismounting period. The mass difference (w 2 -w 1 ) of the dried blank filter paper before (w 1 ) and after the exposure period (w 2 ) was evaluated and subtracted from the sample mass.
The road dust samples were collected using a stainless-steel scoop from eight locations along Raipur city's highway: Tatibandh, Hirapur, Sarora, Khamtarai, Birgaon, Urla, Sankra, and Siltara, during May 2008. These examined sites were situated in an area of approximate 10 km radius. The last four sites listed were located in the industrial area. The Khamtarai site was located in a heavy traffic area. The sample was collected from both sides (left and right) of the road junction in two 250-mL glass bottles. These samples were mixed in equal mass ratio to form a composite sample. Similarly, the sampling was extended up to period 2015, and one sample was taken every year in May from 2009 to 2015 at the Khamtarai site for the temporal variation studies. The dust samples were sundried for one week, and further dried at 50°C in a hot air oven overnight. The samples were then crushed into fine-powder form by sieving out of particles of mesh size >100 µm and finally stored in an aluminum foil for the analyses.
Ten surface sediment samples (0-10 cm depth) were collected in May 2008 from 10 ponds (Rohnipuram, Ashi, Budheshwar, Raja, Pandri, Siltara, Urkura, Birgoan, and Sarora) of Raipur city. These ponds were situated over an area of approximately 25 km 2 . The last five listed ponds were located at industrial sites. For depth studies, two more sediment samples at a depth of 10-20 and 20-30 cm were collected in Siltara in 2008. The Raja pond was chosen for the temporal evolution investigation, and one sample was collected every year in May from 2009 to 2015.
particulates with the snow in the winter season (Ganguly et al. 2019).
Polychlorinated biphenyls are considered as significant environmental air contaminants due to their ecotoxicological and human health implications. The PM in the air is associated with large fraction of EC, OC, and CC. The concentration of the PM 10 , EC, OC, and CC in the ambient air of Raipur city monitored at four locations (Kota, Pt. R. S. University, Raipura Chowk, and Amapara) were in the 388-845, 33-81, 23-52, and 28-65 µg m −3 range, respectively, with average values of 574 ± 222, 53 ± 23, 36 ± 14, and 44 ± 18 µg m −3 , respectively. The highest PM 10 and carbon concentrations in the air corresponded to samples collected in Amapara, most probably due to vehicular emissions.
The concentrations of the PCBs congeners in the ambient air associated to the particulates (PM 10 ) are presented in Table 1. A total 83 congeners in the ambient air were detected. The total concentration of PCBs (n = 4) ranged from 72.07 to 429.57 pg m −3 , with a mean value of 247.54 pg m −3 . The maximum PCBs value was found for the Kota site, expected due to vehicular emissions. The dominant congeners (1, 3, and 17) were found at concentrations in the 2. 4-130.41, 49.31-147.74, and 1.42-18.8 pg m −3 ranges, with mean values of 46.97, 96.10, and 9.01 pg m −3 , respectively.

Meteorology and Concentrations of Particulates, Carbon, and PCBs in Air
The meteorology (i.e., temperature, humidity, vapor pressure, wind speed, and sunshine) for the study period February 2007 to January 2008 ranged from 19.7°C to 38.1°C, 20.8% to 92.8%, 7.1 to 24.5 mm, 1.5 to 12 km h −1 , and 0.9 to 10 h day −1 , respectively. The lowest values of the ambient temperature, vapor pressure, wind speed, and sunshine were observed from December to January. A total rainfall of 96 cm, together with the maximum humidity, was registered from July to September 2007. The winds blew from the northeast and showed speeds of approximately 11 km during the June to August 2007 period. The lowest wind speed from the N direction was observed during the October 2007 to January 2008 period.
The major fraction of the PM was composed of carbons: EC, OC, and CC. The concentration of (n = 24) PM 10 ,EC 10 ,OC 10 , and CC 10 in the air during the February 2007 to January 2008 period ranged from 116 to 523, from 8.8 to 65.5, from 7.2 to 55.4, and from 6.1 to 58.7 µg m −3 , with mean values of 283 ± 138, 28.8 ± 16.9, 23.2 ± 13.6, and 22.7 ± 16.9 µg m −3 , respectively. The PM showed a negative correlation with the meteorological factors (i.e., rain fall, temperature, humidity, vapor pressure, and wind speed) of r = −0.31-0.74. Three meteorological factors (rain, wind speed, and wind direction) markedly influenced the PM concentration. The highest mass concentration was observed in the winter season, December-January, expected due to the lowest atmospheric pressure, temperature, and wind speed. In turn in the rainy season (June-August), the lowest concentration was recorded, probably due to PM washout with the rain.
A similar distribution trend of the PM and carbons in the Raipur city was reported by Jaiswal et al. (2019b). In the hilly area of north India, the different distribution trend of the PM and carbons (maximal in the summer season) was likely due to precipitation of the  Table 2. The total concentration of PCBs ranged from 102 to 537 ng g −1 , with an average value of 241 ng g −1 .

Concentration of Carbon and PCBs in Particulate, Dust, and Sediments
EC, OC, and PCBs are emitted during various combustion and industrial processes (Brunciak et al. 2001) and are distributed in various environmental compartments (viz. air, water, dust, and soil) in urban and industrial areas (Malina and Mazlova 2017). Along roads, they are predominantly emitted by vehicular emissions . The environmental PCBs are transported to water reservoirs by rain, runoff water, industrial and municipal waste, etc. (Froese et al. 1997). Tables 1-3 summarize the environmental contamination of PCBs in the ambient particulates, road dust, and sediments of the study area.
The concentration of EC, OC, and CC in the coarse particulate matter (PM 10 ) ranged from 8.4% to 9.61%, from 5.45% to 8.81%, and from 7.54% to 9.43%, respectively, with mean values of 9.15% ± 0.55%, 6.87% ± 1.20%, and 8.74% ± 0.61%, respectively. The comparable EC concentration in the road dust and sediments varied from 5.8% to 6.61% and from 7.29% to 7.77%, with average values of 6.26% ± 0.31% and 7.56% ± 0.21%. However, very low OC (0.39% ± 0.09% and 0.49% ± 0.06%) and CC (0.15% ± 0.03% and 0.12% ± 0.02%) concentrations in the dust and sediment samples were found, as compared to the particulate samples. A higher EC concentration in the studied area than in other locations reported in the literature, both in India and in the rest of world, was observed (Zong et al. 2016;Guha et al., 2015;Han et al. 2015;Ozdemir et al. 2014;Han et al. 2009). This may be tentatively ascribed to massive coal burning in the area, given that two of India's largest coal-fired power stations are operating in Chhattisgarh state.
The concentration of the individual PCBs congeners in the ambient air and particulates is presented in Table 1. The total concentration of PCBs in the ambient particulate matters of four locations ranged from 185.72 to 1,110.82 ng · g −1 , with an average P r e p r i n t P r e p r i n t vehicular and industrial emissions (Fig. 4). These data represented an average annual increment rate of 6.2% ± 3.2%, 4.9% ± 2.1%, and 5.4% ± 1.1% in PCBs concentrations, respectively.
The main PCBs emission sources are the chlorination of biphenyls during the combustion process of fuels and other materials, vaporization; leakage from application sites of Aroclors; and burning, disposal, and dumping of PCBs containing materials (Meijer et al. 2003).
The combustion of fuels, industrial/metallurgical activities and power generation processes have been reported as possible sources of PCBs, due to reaction of carbon and chlorine at the combustion source (Biterna and Voutsa 2005;Dyke et al. 2003;Weber et al. 2001).
The total PCBs content showed a fair correlation (at r = 0.73-0.75) with PM, EC, OC, and CC, indicating that PCBs would be partly originated from burning processes (Table 4). MCBs, DCBs, and TCBs had a good correlation with TCBs, TeCBs, HeCBs, and OCBs (r = 0.79-0.93); TeCBs, PeCBs, HCBs, HeCBs, and NCBs (r = 0.76-0.91); and MCBs, HeCBs, and OCBs (r = 0.91-0.96), respectively, indicating that they would be originated by the chlorination process of lower congeners (1b)-(1g) as follows (Biterna and Voutsa 2005;Dyke et al. 2003;Weber et al. 2001) C 12 H 10 + Cl 2 C 12 H 9 Cl + HCl (1b) C 12 H 9 Cl + Cl 2 C 12 H 8 Cl 2 + HCl (1c) C 12 H 8 Cl 2 + Cl 2 C 12 H 7 Cl 3 + HCl (1d) When the PCBs concentration in the pond sediment samples was analyzed as a function of depth, it was observed that the concentrations of MCBs, DCBs, TCBs, TeCBs, PeCBs, and HeCBs congeners increased steeply with depth, probably due to poor adsorption on the sediment particles (Fig. 2). However, the concentration of HCB decreased as the depth increased, suggesting an adsorption by the top-layer sedimentary particles (Fig. 2). A noticeable vertical distribution of congeners 1, 3, 4 + 10, 6, 7 + 9, 8 + 5, 17, 18, 19, 77 + 110, and 85 was observed. For comparison purposes, the PCBs concentration in six ponds in Bhilai and Korba varied from 201 to 648 ng · g −1 and from 404 to 773 ng g −1 , with average values of 480 ± 150 and 561 ± 155 ng g −1 , respectively (Patel et al. 2013). The concentrations of aforementioned congeners in these two cities (in which huge quantities of coal are burnt for steel and electricity production) were even higher than those found in Raipur. It is also worth noting that the concentration of PCBs (328 ± 99 ng g −1 ) in the pond sediments from the area of study was higher than those observed (<0.01-126.49 ng g −1 ) in soil/sediment of other locations Li et al. 2012;Kumar et al. 2011), except for Harbor Island East, with concentrations of up to at 1,387 ng g −1 .

Comparison of PCBs Concentrations in Environmental Samples
Comparable total PCBs concentrations were observed in the pond sediments and road dust samples, while their concentration in the particulate samples was markedly increased [ Fig. 3(a)]. The highest concentrations of MCBs, HeCBs, OCBs, and NCBs were found in the particulates; those of TeCBs and PeCBs in the road dusts (Fig. 3); and those of DCBs, TCBs, and HCBs in the sediments (Fig. 3). Remarkably higher contents of MCBs and HeCBs were detected in the PM samples, which could be an indicator of air pollution.
A high concentration of congeners 3, 17, 89, and 138 + 158 was identified in the PM and sediment samples, indicating emissions by multiple sources. The dominant concentration of congeners 1, 8 + 5, 49, and 91 was registered in the road dust samples, showing emissions mainly by vehicles.
As per the ANOVA test, the uncertainty (F) value for the PCBs concentration in the particulates, road dust, and sediments were found to be 7, 65, 535, and 29, indicating multiple emission sources of the PCBs in the road geo-media.
With regard to the total concentrations of PCBs in the dust, electronic waste, and sediments reported in other locations: Guangzhou and Hong Kong, Chennai, Northern Vietnam, Nakdong River (Korea), San Diego Bay, China, and Delhi , it ranged from 4.02 to 114 ng kg −1 , from 1.6 to 53, from 0.25 to 14, from 0.124 to 79.2, from 23 to 1,387, and from <0.01 to 99.40 ng g −1 , respectively. Again, the values detected in this study would be among the highest reported in the literature.
Thus, the values found for Raipur would be higher than those found in Korea, China, and Delhi, and would only be exceeded by those detected in San Diego Bay, California.  These findings corroborate those reported by Chakraborty et al. (2013) in a study on atmospheric PCBs levels (gaseous and particulate phase) in the Indian cities: New Delhi, Agra, Kolkata, Mumbai, Goa, Chennai, and Bangalore. In view of the detected concentrations, the authors urge increasing control over the release of PCB sources in India and ask for measures to protect human health and the environment.

Conclusions
Very high concentrations of PCBs (mainly congeners 1 and 3) were detected in Raipur area: 143 pg m −3 (365 ng g −1 ), 194, and 299 ng g −1 for particulate matter, road dust, and sediments, respectively. These contents are among the highest ever reported in the literature, and clearly exceeded the recommended value of 60 ng g −1 . The normalized total mean TEFs for the PM, road dust, and sediments from Raipur city were estimated at 0.00066, 0.00058, and 0.00047, respectively. In view of the correlations with particulate matter, elemental carbon, organic carbon, and carbonate carbon, massive coal burning and vehicular emissions in Chhattisgarh region can be ascribed as the main sources of PCB pollution. Temporal evolution, tracked over an eight-year period, showed an average annual increment rate of approximately 5.4% in PCBs concentration, while vertical profile analyses showed substantial PCBs concentrations at deeper sediments. Industrial uses, and coal and biomass combustion were apportioned as the major sources of PCBs contamination in the studied area. The collected data points to a dramatic situation, which calls for urgent action to meet the Stockholm Convention goals.

Data Availability Statement
All data, models, and code generated or used during the study appear in the published article.