Environmental Pollution 211 (2016) 226e232

Contents lists available at ScienceDirect

Environmental Pollution

journal homepage: www.elsevier.com/locate/envpol

Occurrence, distribution and source apportionment of polychlorinated naphthalenes (PCNs) in sediments and soils from the Liaohe River Basin, *

Fang Li a, b, Jing Jin a, Yuan Gao a, Ningbo Geng a, b, Dongqin Tan a, b, Haijun Zhang a, * Yuwen Ni a, Jiping Chen a, a Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, No. 457 Zhongshan Road, Dalian 116023, China b University of Chinese Academy of Sciences, Beijing 100049, China article info abstract

Article history: The occurrence and spatial distribution of polychlorinated naphthalenes (PCNs) were investigated in Received 23 July 2015 sediments, upland and paddy soils from the Liaohe River Basin. Concentrations of SPCNs were in the Received in revised form range of 0.33e12.49 ng g 1 dry weight (dw) in sediments and 0.61e6.60 ng g 1 dw in soils, respectively. 28 September 2015 Tri-CNs and tetra-CNs were the dominating homologues. An increasing trend of PCNs contamination was Accepted 29 September 2015 found in sediments with the rivers flowing through industrial areas and cities. Soils collected near cities Available online 14 January 2016 exhibited higher abundance of PCNs than that of rural areas. The distribution of PCNs was related to the local industrial activities, rather than total organic carbon. Positive matrix factorization (PMF) was used Keywords: Polychlorinated naphthalenes for the source apportionment of PCNs in sediments and paddy soils. The result of PMF indicated that Occurrence PCNs in sediments and paddy soils were mainly from the industrial processes, with additional contri- Distributions butions from the historical use of Halowax 1014 and atmospheric deposition. Source apportionment © 2015 Elsevier Ltd. All rights reserved.

1. Introduction been ubiquitously detected in numerous environmental and bio- logical samples (Bidleman et al., 2010a). The occurrence of PCNs in Polychlorinated naphthalenes (PCNs), a family of chlorinated sediments and soils was in the low to sub-ppb concentrations polycyclic aromatic hydrocarbons, consist of 75 possible congeners. (Castells et al., 2008; Helm et al., 2008a; Ishaq et al., 2009; Meijer They were commercially produced and mainly used in the electrical et al., 2001b; Schuhmacher et al., 2004). Currently, only limited industry with ideal chemical properties and thermal stability in the studies have focused on PCNs contaminant in China (Pan et al., 20th century (Falandysz, 1998). It has been identified that some 2011; Wang et al., 2012; Xu et al., 2015). The Liaohe River is one congeners can induce dioxin-like toxicity via binding to the aryl of the most heavily polluted rivers in China, and runs through many hydrocarbon receptor in vivo (Domingo, 2004). Due to their po- large industrial areas of Province. It consists of the Hun tential properties of persistent organic pollutants (POPs) and River, Taizi River and . The Hun River converges with adverse effects on living organisms including humans, PCNs have the Taizi River and flows into the Daliao River before entering into been listed in Annexes A and C of the Stockholm Convention on the Bo Sea. The Liaohe River serves as an important resource for POPs in May 2015 (Stockholm Convention). drinking water, aquaculture and industrial use. Moreover, the Sediments and soils hold the bulk of contemporary POPs car- Liaohe River Basin is also an important agricultural area, and rice is rying aquatic and terrestrial environmental burden. Thus, it is the major agricultural crop. There are more than one million acres essential to monitor POPs in sediments and soils which were of paddy fields irrigated with river water. In 2011, our research approved as secondary pollution sources. As reported, PCNs have group investigated the levels of PCNs in sediments from the Daliao River Estuary with the range of 0.033e0.284 ng g 1 dw, and it was found that organic matter and molecular properties could influence * This paper has been recommended for acceptance by Jay Gan. the partition behavior between sediment and water phase (Zhao * Corresponding author. et al., 2011). However, further detailed studies are needed to E-mail address: [email protected] (J. Chen). http://dx.doi.org/10.1016/j.envpol.2015.09.055 0269-7491/© 2015 Elsevier Ltd. All rights reserved. F. Li et al. / Environmental Pollution 211 (2016) 226e232 227

13 investigate the PCNs pollution in this basin for environmental solutions (2.0 ng, C10eCN 27, 42, 52, 67, 73 and 75) were added to assessment. subsamples as the surrogate standard to monitor the analytical Although their production was prohibited in most countries in recovery efficiency. Ten gram subsamples were soxhlet-extracted 1980s, PCNs can also be released into the environments via unin- with n-hexane/acetone (1:1, v/v) for 16 h with copper granules tentionally emissions from the industrial thermal processes and (2.0 g) to remove the sulfur, and the extract was concentrated to other chlorination processes (Liu et al., 2014). Several approaches approximately 1 mL with rotary evaporator. Then gel permeation have been used to identify sources of PCNs in the environmental chromatography (GPC) with SX-3 Bio-Beads column was used to matrices. Ratios or fractions of several indicator congeners of PCNs remove lipid and other high molecular weight interferences using such as CN 36/45, 42, 54, 53/55, 66/67, 71/72 have been widely used dichloromethane as a mobile phase at the flow rate of 5 mL min 1. to assess the contribution of industrial thermal processes (Liu et al., The extracts were passed through an open silica column packed 2014, Liu et al., 2012). However, overlaps of different industrial with 5.0 g 5% deactivated silica gels (activated at 650 C for 5 h and thermal sources limit its use in source identification of PCNs. Prin- then deactivated with 5% (w/w) Milli-Q water). PCNs were eluted cipal component analysis (PCA) has also been used to examine the with 80 mL of n-hexane, then the fraction was reduced in volume indicator congeners of PCNs based on the relationship between the and further cleaned up on an alumina column (from bottom to top: individual congener profiles of various possible sources (Bidleman 2.0 g anhydrous sodium sulfate, 10.25 g alumina, 2.0 g anhydrous et al., 2010b; Liu et al., 2014). Nevertheless, it is difficult to assess sodium sulfate). The eluent was discarded with 40 mL of n-hexane, the relative importance of individual sources. Currently, the chemical then PCNs were eluted with 120 mL of 5% dichloromethane in n- mass balance (CMB) and positive matrix factorization (PMF) are two hexane. Finally, the final eluent was concentrated and exchanged of the widely-used multivariate receptor modeling approaches for into 200 mL of nonane containing d7-CN 2 (2.0 ng) for instrumental the source apportionment of pollutants in the environments (Seike analysis. PCNs were analyzed by a gas chromatography-triple et al., 2007; Watson et al., 2001). The main shortcoming of the quadrupole mass spectrometer (GCeMS/MS, ThermoFisher Scien- CMB model is that it requires information of all sources and their tific) with electron impact source (Li et al., 2014). Details on the profiles before beginning the analysis. Whereas PMF model can operation parameters of chromatography and MS are presented in reconstruct the profile of potential sources through decomposing the Table S2 of the SI. sample data matrix consisting of sample concentration and uncer- tainty into factor profiles and factor contributions. Then the potential 2.3. Quality assurance and quality control emission sources would be identified based on fingerprints or composition of sources. Hitherto, few studies have applied the PMF An internal standard isotope-dilution method was employed for model for the source apportionment of polychlorinated dibenzo-p- the quantification of PCNs in the current study. It was assumed that dioxins, polychlorinated dibenzofurans (PCDD/Fs) (Sundqvist et al., the response of individual congeners of each homologue group was 2010), polychlorinated biphenyls (PCBs) (Praipipat et al., 2013)in equal to that of the 13C-labled congeners with the same degree of sediments. However, the PMF model has not been mentioned for the chlorination except for congeners of tri-CNs. For tri-CNs congeners, 13 source apportionment of PCNs in environmental matrices. the response of C10eCN 42 was used. The quantification of PCNs The objective of the present study was to investigate the total was performed by using a relative response factor (native to concentrations (SPCNs, tri-to octa-CNs) and spatial distribution of labeled) of the labeled congener at the same level of chlorination PCNs in sediments and ambient upland and paddy soils from the and the similar retention time. The concentration of each homo- Liaohe River Basin. Then the PMF model was used to generate logue is equal to the sum of the concentration of all congeners in candidate source profiles and estimate their relative contribution to this homologue. The limits of detection (LODs) and limits of PCNs in sediments and paddy soils. It is hoped that the work will be quantification (LOQs) of PCNs congeners were governed by the S/N helpful for creating the inventories, understanding sources and ratio. LODs were defined as the minimum amount of PCNs conge- environmental fate and further developing treatment strategies for ners with three times S/N in a chromatogram, and LOQs were controlling PCNs contamination in China. determined as ten times of S/N in blank samples. The LOQs ranged from 0.10 to 1.75 pg g 1 dw (dry weight) for individual congeners of 2. Material and methods PCNs (Table S2). Surrogate recoveries of PCNs were within the range of 63%e146%. The procedural blank samples and duplicate 2.1. Sample collection samples were processed for every ten samples. No congeners were detected in the procedural blank. Forty-five samples (22 sediments, 5 upland soils and 18 paddy soils) were collected from the Liaohe River Basin in June 2010. 2.4. Data analysis Sediment and soil samples (top 0e5 cm) were taken with a stain- less steel grab sampler. Each sampling site was represented by five All statistical analysis was performed with SPSS software subsamples from the area of 25 25 m. The distributions and (version 18.0, SPSS Inc., Chicago, IL, USA). In all instances, concen- detailed location information of sampling sites are shown in Fig. 1, tration was assumed to equal half the LOQs when the concentration Fig. 2 and Table S1 of supporting information (SI). The collected was below the LOQs. Spearman correlation analysis was applied to samples were transported to the laboratory on ice, then freeze- assess the relationship between concentration of individual PCNs dried, ground and homogenized by sieving through a stainless homologues (tri-to octa-CNs) and TOC. Changes of PCNs congener 60-mesh sieve. All samples were stored in amber glass bottles at patterns have been unknown and they have not been estimated in below 4 C until further analysis. The total organic carbon (TOC) in transportation processes. Therefore, the PMF model is used to sediment and soil samples was analyzed by an elemental analyzer apportion sources of PCNs based on the assumption that PCNs from (elementar Vario Macro CHN, Germany), and the results are also different sources are linearly additive and no degradation or other presented in Table S1 of the SI. pattern alteration takes place during the transportation process from the sources to receptor. The PMF software (version 5.0, U.S. 2.2. Preparation and analysis Environmental Protection Agency) was used in this study. Forty- five peaks (63 PCN congeners) were pared down to 28 peaks rep- 13 Prior to extraction, the C10-isotopic PCN mixed standard resenting 39 congeners in the original 22 sediment samples and 18 228 F. Li et al. / Environmental Pollution 211 (2016) 226e232

paddy soil samples. A distinct advantage of PMF is its ability to Qinghai Plateau (0.013e0.029 ng g 1 dw) (Pan et al., 2013), the handle incomplete data such as below the detection limits, missing Pearl (0.0095e0.666 ng g 1 dw) (Wang et al., 2012) and data and outliers by adjusting the error estimates of each data the forest soil (0.045e0.140 ng g 1 dw) (Xu et al., 2015) in China. point. The uncertainties for values above the LOQ were substituted with the analytical uncertainty plus one-thirds of the LOQ, 10% of 3.2. Spatial distributions of PCNs the measured concentrations were estimated as analytical un- certainties. Values below the LOQ were replaced by half of the LOQ, As shown in Fig. 1, the levels of PCNs in sediments from the Hun fi and uncertainties were substituted with ve sixths of the LOQ. River and Taizi River increased with flowing through industrial Missing values were replaced by the geometric mean of the areas and living districts. PCNs concentration of sample S1 collected measured values and their accompanying uncertainties were set at in the upper reaches of the Hun River was 0.70 ng g 1 dw, indi- four times of this geometric mean value. This treatment of un- cating that such low concentration might be considered as the certainties was in line with recommendations from Kim et al. (Kim environmental background level. It was possible that this site was et al., 2005). A bootstrapping analysis of the base run with the located in the downstream channel of the Dahuofang Reservoir and lowest objective function (Q) values could be conducted to identify was less influenced by various industrial activities. The PCNs con- the correct number of factors in PMF 5.0 software. Three to ten centration of sample S4 collected near the Fushun Municipal fi factor solutions were generated. The coef cient of determination Wastewater Treatment Plant (MWTP) was 4.90 ng g 1 dw, much 2 fi (COD, r ), Q and residuals were used to evaluate the goodness of t greater than sample S3 (upstream of the MWTP). High PCNs con- between the modeled data set and the actual data set. The COD centration could be ascribed to the discharge of effluents from describes the ability of the model to reproduce the original data and MWTP. Similarly, sample S6 also had a higher PCNs concentration fi equals 1.0 for a perfect t. The ratio of Q (True) to Q (Robust) was of 7.02 ng g 1 dw. It might be attributed to the partial industrial fi below 1.5 indicates a good t. Finally, the residuals are symmetri- sewage of Shenyang flowing into the Hun River at upstream of site cally distributed with absolute values below 3 in a good model. S6. Additionally, samples S8eS11 situated the lower reaches of the More detailed information on PMF model was described in the Hun River showed high PCNs concentrations, ranging from 3.40 to supporting information. 12.25 ng g 1 dw. It was reasonable that the presence of the steel and iron industries in city probably aggravated the PCNs 3. Results and discussion pollution in the study area. PCNs concentration in the lower reaches of the Taizi River was also high, with concentrations ranging from 3.1. Levels of PCNs in sediments and soils 4.16 to 12.49 ng g 1 dw. In the upper and middle reaches of the Daliao River, the concentrations of PCNs were 1.56 ng g 1 dw and Concentrations of PCNs in sediment and soil samples are sum- 2.80 ng g 1 dw, respectively. In addition, the low concentrations of marized in Table S3 in terms of concentration range, mean and SPCNs in sediments from Daliao River Estuary could be related to medium values. Congeners including CN 18, 25/26, 29, 31, 32, 39, 41, the dilution effect of seawater (Zhao et al., 2011). 44, 51, 56 and 70 were not detected in all sediment samples. The In soils, there was significant discrimination in spatial distri- concentrations of SPCNs in sediments were in the range of bution of PCNs (see Fig. 2). The soil samples collected near She- 0.33e12.49 ng g 1 dw, with an average concentration of nyang City had relatively high PCNs concentrations ± 1 4.42 3.63 ng g dw. The mean concentration of tri-CNs and octa- (3.08e6.19 ng g 1 dw), and such concentrations were higher than CNs had higher than other homologues in three rivers. Based on the those in rural areas. The result demonstrated that the urban sources relative potency factors (RPFs) (Noma et al., 2004), the calculated had a remarkable contribution to the PCNs contamination in soils. PCN-corresponding total toxic equivalents (STEQs) in sediments 1 Concentrations of PCNs in paddy soils collected near Shenyang City ranged from 0.02 to 4.86 pg TEQ g . Concentrations of PCNs in were corresponding with those in upland soils without irrigation, sediments were similar to these studies reported other areas, such indicating that irrigation might be not the major pathway of PCNs as the Gulf of Bothnia (0.27e2.8 ng g 1 dw) (Lundgren et al., 2003) in paddy soils. In other words, it was suggested that atmospheric and the Laizhou Bay area in China (0.12e5.1 ng g 1 dw) (Pan et al., 2011). However, the result was greatly lower than those reported in the Lake Ontario (21e38 ng g 1 dw) (Helm et al., 2008b). The site- specific potential risks of exposure to PCNs in sediments were assessed, and up to 45% and 27% of sediments exceeded the interim sediment and quality guidelines (ISQGs) in Canada and USA (0.85 pg TEQ g 1 dw and 2.5 pg TEQ g 1 dw) (Zhao et al., 2011), respectively. The result showed that these sediments may have a risk of deleterious effects on the sensitive organism at all trophic levels. However, there was no any site above the probable effect levels (PLEs) guidelines in Canada and USA (21.5 pg TEQ g 1 dw and 25 pg TEQ g 1 dw) in our study area. CN 18 was detected in some soils, whereas CN 9 and CN 10 were not detected in all soils. The concentrations of SPCNs in upland and paddy soil samples ranged from 0.95 to 3.55 ng g 1 dw and 0.61e6.60 ng g 1 dw with a mean value of 2.84 ng g 1 dw and 2.37 ng g 1 dw, respectively. Mean concentration of tetra-CNs was the highest in upland soils, whereas that of tri-CNs was the highest in paddy soils. The PCN-corresponding STEQs in upland and paddy soils were in the range of 0.13e1.18 pg TEQ g 1 and 0.06e2.58 pg TEQ g 1, respectively. The result in soils was significantly lower e 1 than that of the River Chenab in Pakistan (24.6 223.0 ng g dw) Fig. 1. Spatial distribution of PCNs in sediments from the Liaohe River Basin. Date of (Mahmood et al., 2014), but higher than that of the eastern Tibei- samples S23eS28 taken from reference reported by Zhao et al., 2011. F. Li et al. / Environmental Pollution 211 (2016) 226e232 229

2007). The derived ratio in our study ranged from 0.05 to 0.23 (Table S5), indicating that both of combustion-related sources and re-emission of Halowax products influenced PCNs pollution in the Liaohe River Basin. Additionally, CN24 was detected in all samples as an indicator of coal/wood burning and cooking, implying that the coal/wood combustion was also sources of PCNs in the Liaohe River Basin (Dong et al., 2013; Lee et al., 2000).

3.4. Source apportionment

Four factors were resolved from the data sets of sediments and paddy soils, respectively. Criteria used to determine the correct number of factors is described in the supporting information. We identified resolved factors of PCNs by weight evidence based on their congener patterns matching with those of real-sources re- ported in articles. The profiles of the resolved PCNs sources (factors) in sediments generated by the PMF software are shown in Fig. 3. The PMF model resulted in one source profile is dominated by tri-CNs (Factor 1). The second factor is characterized as CN 33/34/37, 27/30 and 50 Fig. 2. Spatial distributions of PCNs in upland soils and paddy soils from the Liaohe River Basin. MWTP ¼ Municipal Wastewater Treatment Plant. with low contribution from tri- and hexa-CNs (Factor 2). Factor 3 consists of CN 55/53 and CN 59 but with low contribution from tetra-CNs. The last factor is dominated by high chlorinated naph- deposition might be a major source of PCNs in upland and paddy thalenes (hexa-to octa-CNs). They account for 20.0%, 10.0%, 20.6% soils. The conclusion was in agreement with that reported by Pan and 48.6% of the total PCNs in the data set of sediments from the et al. (Pan et al., 2013). Those paddy soils collected from nearby the Liaohe River Basin, respectively. upper and middle reaches of the Daliao River also had relatively low Tri-CNs is predominant homologue in the air and deposition PCNs contamination since there were few industrial activities samples in China and other areas/regions (Hogarh et al., 2012; Lee nearby. et al., 2007; Li et al., 2012). However, high abundance of low The correlations between the concentrations of individual PCNs chlorinated homologues in atmosphere was also linked to indus- homologues and TOC were statistically analyzed. The insignificant trial thermal processes (Lin et al., 2013; Liu et al., 2012). Emission of correlation between them suggested that local emission sources PCNs from industrial thermal processes does not always follow a play a more crucial role than TOC in the distribution of PCNs in standard pattern since the formation of PCNs is dependent on sediments and soils from the Liaohe River Basin (Table S4). Mean- factors such as process designs, temperatures, raw materials. Liu while, it was also found that all PCNs homologues were closely et al. reported that low chlorinated homologues (di-to penta-CNs) related to each other in sediments and soils, except for tri-CNs. dominated the total concentration of PCNs in most of the stack gas samples from industrial thermal processes such as iron steel 3.3. Compositional profiles of PCNs plant, municipal solid waste incinerator (MSWI) and nonferrous smelting (Liu et al., 2015). Some congeners including CN 33/34/37, The homologue profiles of PCNs in sediments and soils are 27/30, 45/36, 51, 50, 52/60 and 66/67 were abundant in these flue shown in Fig. S1, and each sample displayed a unique homologue gases. The vapor pressures of low chlorinated homologues have profile that was different from those of Halowax series (Li et al., higher vapor pressures than other high chlorinated homologue, 2014). Low chlorinated homologues comprising of tri-CNs and especially tri-CNs. Tri-CNs easily tends to long-range transport in tetra-CNs were dominant in most samples, accounting for above the atmosphere with the estimated vapor pressures to be 0.13 Pa 50% of total PCNs, followed by octa-, penta/hexa- and hepta-CNs. As (25 C). Thus we hypothesize that Factor 1 and Factor 2 represent shown in Fig. S1-a, obvious variations in the homologues profiles the contribution from atmospheric deposition to sediments in this were observed in sediments from different reaches of the rivers. basin, with reflecting global background air (I) and regional in- The abundance of high chlorinated homologues in sediments ten- dustrial emission (II), respectively. In China, PCNs mixtures were ded to increase in the lower reaches of the Hun River and Taizi never historically produced and are not currently in commercial River, indicating that an important PCNs source was present with use. It is somewhat surprising that the profile of Factor 3 closely high chlorinated homologues as the dominating homologues. resembles that of PCNs in Halowax 1014 (Falandysz et al., 2006), However, it was notably different that lower chlorinated homo- indicating Factor 3 may represent the evaporative emissions from logues were the dominating homologues in ambient paddy soils the past use of Halowax 1014. This is not surprising given that there (Fig. S1-b). Tri-CNs was the predominant homologue of PCNs in are some electrical industries and e-waste recycling industries in paddy soils, especially for paddy soils collected from the rural this study area. None of the real sources fingerprints match with nearby the Daliao River. In upland soils, the dominant homologues Factor 4. Fly ash is considered to be one of the most matrices of PCNs in sample U1eU3 were tri-CNs and tetra-CNs, which was emitted from industrial thermal processes, and it can get into the different from that of sample U4 and U5 (octa-CNs). river with wastewater and then undergo sedimentation. The As reported, the contribution of SPCNscom (CN 17, 25/26, 27/30, congener profiles of PCNs were different from that of fly ash sam- 35/48, 36/45, 39, 50, 51, 52/60, 54, 66/67) to SPCNs (tri-to octa-CNs) ples emitted from the different thermal processes (Meijer et al., can be used to estimate the contribution of combustion-related 2001a; Nie et al., 2011). Several studies showed that these conge- congeners, the ratio>0.5 was identified as the combustion-related ners including CN 66/67, 73 and 75 were dominant in fly ash sources (iron-steel production, metal refining, waste incineration samples from MSWI (Takasuga et al., 2004) and secondary and coal, wood and biomass combustion), while that <0.11 was nonferrous metallurgical facilities (Ba et al., 2010; Nie et al., 2012). considered as the re-emission of Halowax products (Lee et al., The pattern within each homologue of PCNs from the same 230 F. Li et al. / Environmental Pollution 211 (2016) 226e232

Fig. 3. Congener profiles obtained when applying PMF to PCNs data of sediments from the Liaohe River Basin assuming four sources (factors). PCN congener numbers on the x-axis * are plotted versus their percent of each congener of PCNs on the y-axis. (CN 63 : excluded in the data set of sediments, but included in that of soils). industrial thermal process is often similar. In Fig. S2 of the and the fraction contribution was high for sample P11 and P18. supporting information, Factor 4 is overall in agreement with the However, a high fraction contribution of atmospheric regional PCNs isomer profiles of several industrial thermal processes. deposition was found in sample P2 with the maximum concen- Furthermore, the dominating congeners in the graphite sludge tration (4.61 ng g 1 dw). The maximum concentration of the past from chlor-alkali plant were CN 52/60, 66/67 and 73 (Jarnberg€ et al., use of Halowax 1014 was 1.94 ng g 1 dw, and the fraction contri- 1997; Kannan et al., 1998). Therefore, we speculate that Factor 4 bution was just high for sample P4 and P14. represents industrial thermal processes and other chlorination This study suggests that the local industrial processes are the processes in Liaoning Province. most important source of sediments and soil from the Liaohe River The calculated relative source contributions and spatial distri- Basin. The result was different from that PCNs in sediments mainly bution for sediment samples are illustrated in Fig. 4. Atmospheric attributed to the industrial use of Halowax 1014 in the Laizhou Bay background deposition concentrations ranged from not detected area, North China (Pan et al., 2011). This is probably related to the (nd) to 2.24 ng g 1 dw, and the fraction contribution was great in industrial dissimilar patterns in the two regions. Due to the overlap the upper reaches of the Hun River and the Taizi River, such as of the sources profile (Liu et al., 2014), it was difficult to attribute samples S1 and S12. Industrial process exhibited variable concen- the PCNs contamination in this basin to a specific industry source. trations ranging from nd to 6.29 ng g 1 dw. The relative contribu- Furthermore, preferential degradation of PCNs may occur under tion of Factor 4 to the total PCNs was high in sediments from the sunlight conditions in the atmospheric transport and deposition, middle and low reaches of the Hun River and the Taizi River. These while it cannot degrade under aerobic conditions (Jarnberg€ et al., sites also show high contributions of atmospheric regional depo- 1999). However, no decreasing time trend was found for sition with concentrations ranging up to 1.27 ng g 1 dw. A sub- photolysis-indicator congeners including CN 35, 53, 62 and 69 stantial input of the past uses of Halowax 1014 was predicated (Meijer et al., 2001a), suggesting that photolysis degradation is less samples S5 and S14, and the concentrations ranged up to important in this study area. 5.06 ng g 1 dw in sediments. No clear differences were observed between the profiles of PCNs sources (factors) in paddy soils versus sediments (Fig. S3). The in- 4. Conclusions dustrial processes account for the largest share of PCNs in the data set of paddy soils at 40.0%. The historical use of Halowax 1014, at- In the present paper, we investigated the concentrations and mospheric background and regional deposition contributed 12.6%, distribution of PCNs in sediments, upland and paddy soils from the 21.6% and 25.8% of the total PCNs content of paddy soil samples in Liaohe River Basin, China. Levels of PCNs in sediments and soils 1 this basin, respectively. As shown in Fig. S4, atmospheric back- were less than 13.0 ng g dw, and such concentrations were in ground deposition was the major contributor to most paddy soils agreement with many studies in other regions. Tri-CNs and tetra- that were taken from near the Daliao River and the lower reaches of CNs were the dominating homologues in most of the samples. It Taizi River. Other candidate sources were highly localized. The was found that the distribution of PCNs in sediments was different concentrations of industrial processes ranged up to 3.53 ng g 1 dw, from that in soils. The result of correlation analysis suggested that the distribution of PCNs was mainly related to local industrial F. Li et al. / Environmental Pollution 211 (2016) 226e232 231

Fig. 4. Spatial distributions of the resolved four factors in terms of the fraction of total PCNs and concentration (ng g 1 dw) in sediments from the Liaohe River Basin. activities, rather than TOC. Based on the results of source appor- Environmental Protection Foundation for Public welfare Project tionment by PMF model, it was found that PCNs in sediments and (201309030), the Special Fund for Agro-scientific Research in the soils from the Liaohe River Basin were mainly from the emission of Public Interest (201503108) and the National Basic Research Pro- industrial processes. Moreover, the re-emission of technical PCNs gram of China (973 Program, Grant No. 2015CB453100). mixtures and the contribution of atmospheric deposition are also of importance at some sites in this basin. Additionally, coal and wood Appendix A. Supplementary data combustion and cooking source also exist in this basin. Further studies should be conducted on the environmental behaviors and Supplementary data related to this article can be found at http:// human exposure risks levels of PCNs and the formation mecha- dx.doi.org/10.1016/j.envpol.2015.09.055. nisms and reducing ways of PCNs in industrial processes to control and reduce the PCNs contamination in China. References

Acknowledgments Ba, T., Zheng, M.H., Zhang, B., Liu, W.B., Su, G.J., Liu, G.R., Xiao, K., 2010. Estimation and congener-specific characterization of polychlorinated naphthalene emis- sions from secondary nonferrous metallurgical facilities in China. Environ. Sci. This work was supported by the National Natural Science Technol. 44, 2441e2446. Foundation of China (Grant No. 21205117), the Special Bidleman, T.F., Helm, P.A., Braune, B.M., Gabrielsen, G.W., 2010a. Polychlorinated 232 F. Li et al. / Environmental Pollution 211 (2016) 226e232

naphthalenes in polar environments-a review. Sci. Total Environ. 408, Atmospheric emission of polychlorinated naphthalenes from iron ore sintering 2919e2935. processes. Chemosphere 89, 467e472. Bidleman, T.F., Helm, P.A., Braune, B.M., Gabrielsen, G.W., 2010b. Polychlorinated Liu, G.R., Cai, Z.W., Zheng, M.H., 2014. Sources of unintentionally produced poly- naphthalenes in polar environments - a review. Sci. Total Environ. 408, chlorinated naphthalenes. Chemosphere 94, 1e12. 2919e2935. Liu, G.R., Cai, Z.W., Zheng, M.H., Jiang, X.X., Nie, Z.G., Wang, M., 2015. Identification Castells, P., Parera, J., Santos, F.J., Galceran, M.T., 2008. Occurrence of polychlorinated of indicator congeners and evaluation of emission pattern of polychlorinated naphthalenes, polychlorinated biphenyls and short-chain chlorinated paraffins naphthalenes in industrial stack gas emissions by statistical analyses. Chemo- in marine sediments from Barcelona (Spain). Chemosphere 70, 1552e1562. sphere 118, 194e200. Stockholm Convention, http://chm.pops.int/TheConvention/ThePOPs/TheNewPOPs/ Lundgren, K., Tysklind, M., Ishaq, R., Broman, D., van Bavel, B., 2003. Flux estimates tabid/2511/Default.aspx. and sedimentation of polychlorinated naphthalenes in the northern part of the Domingo, J.L., 2004. Polychlorinated naphthalenes in animal aquatic species and Baltic Sea. Environ. Pollut 126, 93e105. human exposure through the diet: a review. J. Chromatogr. A 1054, 327e334. Mahmood, A., Malik, R.N., Li, J., Zhang, G., Jones, K.C., 2014. PCNs (polychlorinated Dong, S.J., Liu, G.R., Zhang, B., Gan, L.R., Zheng, M.H., 2013. Formation of poly- napthalenes): dietary exposure via cereal crops, distribution and screening- chlorinated naphthalenes during the heating of cooking oil in the presence of level risk assessment in wheat, rice, soil and air along two of the high amounts of sucralose. Food Control 32, 1e5. River Chenab. Pak. Sci. Total Environ. 481, 409e417. Falandysz, J., 1998. Polychlorinated naphthalenes: an environmental update. Envi- Meijer, S.N., Harner, T., Helm, P.A., Halsall, C.J., Johnston, A.E., Jones, K.C., 2001a. ron. Pollut. 101, 77e90. Polychlorinated naphthalenes in UK soils: time trends, markers of source, and Falandysz, J., Nose, K., Ishikawa, Y., Lukaszewicz, E., Yamashita, N., Noma, Y., 2006. equilibrium status. Environ. Sci. Technol. 35, 4205e4213. HRGC/HRMS analysis of chloronaphthalenes in several batches of Halowax Meijer, S.N., Harner, T., Helm, P.A., Halsall, C.J., Johnston, A.E., Jones, K.C., 2001b. 1000, 1001, 1013, 1014 and 1099. J. Environ. Sci. Health Part A Toxic Hazard. Polychlorinated naphthalenes in UK soils: time trends, markers of source, and Subst. Environ. Eng. 41, 2237e2255. equilibrium status. Environ. Sci. Technol. 35, 4205e4213. Helm, P.A., Gewurtz, S.B., Whittle, D.M., Marvin, C.H., Fisk, A.T., Tomy, G.T., 2008a. Nie, Z.Q., Zheng, M.H., Liu, W.B., Zhang, B., Liu, G.R., Su, G.J., Lv, P., Xiao, K., 2011. Occurrence and biomagnification of polychlorinated naphthalenes and non- Estimation and characterization of PCDD/Fs, dl-PCBs, PCNs, HxCBz and PeCBz and mono-ortho PCBs in Lake Ontario sediment and biota. Environ. Sci. Technol. emissions from magnesium metallurgy facilities in China. Chemosphere 85, 42, 1024e1031. 1707e1712. Helm, P.A., Gewurtz, S.B., Whittle, D.M., Marvin, C.H., Fisk, A.T., Tomy, G.T., 2008b. Nie, Z.Q., Liu, G.R., Liu, W.B., Zhang, B., Zheng, M.H., 2012. Characterization and Occurrence and biomagnification of polychlorinated naphthalenes and non- quantification of unintentional POP emissions from primary and secondary and mono-ortho PCBs in Lake Ontario sediment and biota. Environ. Sci. Technol. copper metallurgical processes in China. Atmos. Environ. 57, 109e115. 42, 1024e1031. Noma, Y., Yamamoto, T., Sakai, S.I., 2004. Congener-specific composition of poly- Hogarh, J.N., Seike, N., Kobara, Y., Habib, A., Nam, J.J., Lee, J.S., Li, Q.L., Liu, X., Li, J., chlorinated naphthalenes, coplanar PCBs, dibenzo-p-dioxins, and di- Zhang, G., Masunaga, S., 2012. Passive air monitoring of PCBs and PCNs across benzofurans in the halowax series. Environ. Sci. Technol. 38, 1675e1680. East Asia: a comprehensive congener evaluation for source characterization. Pan, X.H., Tang, J.H., Chen, Y.J., Li, J., Zhang, G., 2011. Polychlorinated naphthalenes Chemosphere 86, 718e726. (PCNs) in riverine and marine sediments of the Laizhou Bay area. North China. Ishaq, R., Persson, N.J., Zebuhr, Y., Broman, D., Naes, K., 2009. PCNs, PCDD/Fs, and Environ. Pollut. 159, 3515e3521. Non-orthoPCBs, in water and bottom sediments from the industrialized Nor- Pan, J., Yang, Y.L., Zhu, X.H., Yeung, L.W.Y., Taniyasu, S., Miyake, Y., Falandysz, J., wegian Grenlandsfjords. Environ. Sci. Technol. 43, 3442e3447. Yamashita, N., 2013. Altitudinal distributions of PCDD/Fs, dioxin-like PCBs and Jarnberg,€ U., Asplund, L., deWit, C., Egeback,€ A.L., Wideqvist, U., Jakobsson, E., 1997. PCNs in soil and yak samples from Wolong high mountain area, eastern Tibet- Distribution of polychlorinated naphthalene congeners in environmental and Qinghai Plateau, China. Sci. Total Environ. 444, 102e109. source-related samples. Arch. Environ. Contam. Toxicol. 32, 232e245. Praipipat, P., Rodenburg, L.A., Cavallo, G.J., 2013. Source apportionment of poly- Jarnberg,€ U.G., Asplund, L.T., Egeback,€ A.L., Jansson, B., Unger, M., Wideqvist, U., chlorinated biphenyls in the sediments of the Delaware River. Environ. Sci. 1999. Polychlorinated naphthalene congener profiles in background sediments Technol. 47, 4277e4283. compared to a degraded Halowax 1014 technical mixture. Environ. Sci. Technol. Schuhmacher, M., Nadal, M., Domingo, J.L., 2004. Levels of PCDD/Fs, PCBs, and PCNs 33, 1e6. in soils and vegetation in an area with chemical and petrochemical industries. Kannan, K., Imagawa, T., Blankenship, A.L., Giesy, J.P., 1998. Isomer-specific analysis Environ. Sci. Technol. 38, 1960e1969. and toxic evaluation of polychlorinated naphthalenes in soil, sediment, and Seike, N., Kashiwagi, N., Otani, T., 2007. PCDD/F contamination over time in Japa- biota collected near the site of a former chlor-alkali plant. Environ. Sci. Technol. nese paddy soils. Environ. Sci. Technol. 41, 2210e2215. 32, 2507e2514. Sundqvist, K.L., Tysklind, M., Geladi, P., Hopke, P.K., Wiberg, K., 2010. PCDD/F source Kim, E., Brown, S.G., Hafner, H.R., Hopke, P.K., 2005. Characterization of non- apportionment in the Baltic Sea using positive matrix factorization. Environ. Sci. methane volatile organic compounds sources in Houston during 2001 using Technol. 44, 1690e1697. positive matrix factorization. Atmos. Environ. 39, 5934e5946. Takasuga, T., Inoue, T., Ohi, E., Kumar, K.S., 2004. Formation of polychlorinated Lee, R.G.M., Burnett, V., Harner, T., Jones, K.C., 2000. Short-term temperature- naphthalenes, dibenzo-p-dioxins, dibenzofurans, biphenyls, and organochlo- dependent air-surface exchange and atmospheric concentrations of poly- rine pesticides in thermal processes and their occurrence in ambient air. Arch. chlorinated naphthalenes and organochlorine pesticides. Environ. Sci. Technol. Environ. Contam. Toxicol. 46, 419e431. 34, 393e398. Wang, Y., Cheng, Z.N., Li, J., Luo, C.L., Xu, Y., Li, Q.L., Liu, X., Zhang, G., 2012. Poly- Lee, S.C., Harner, T., Pozo, K., Shoeib, M., Wania, F., Muir, D.C.G., Barrie, L.A., chlorinated naphthalenes (PCNs) in the surface soils of the Delta, Jones, K.C., 2007. Polychlorinated naphthalenes in the global atmospheric South China: distribution, sources, and air-soil exchange. Environ. Pollut. 170, passive sampling (GAPS) study. Environ. Sci. Technol. 41, 2680e2687. 1e7. Li, Q.L., Xu, Y., Li, J., Pan, X.H., Liu, X., Zhang, G., 2012. Levels and spatial distribution Watson, J.G., Chow, J.C., Fujita, E.M., 2001. Review of volatile organic compound of gaseous polychlorinated biphenyls and polychlorinated naphthalenes in the source apportionment by chemical mass balance. Atmos. Environ. 35, air over the northern South China Sea. Atmos. Environ. 56, 228e235. 1567e1584. Li, F., Jin, J., Sun, X.L., Wang, X.L., Li, Y., Shah, S.M., Chen, J.P., 2014. Gas Xu, Y., Li, J., Zheng, Q., Pan, S., Luo, C., Zhu, H., Nizzetto, L., Zhang, G., 2015. Poly- chromatography-triple quadrupole mass spectrometry for the determination of chlorinated naphthalenes (PCNs) in Chinese forest soil: will combustion atmospheric polychlorinated naphthalenes. J. Hazard. Mater. 280, 111e117. become a major source? Environ. Pollut. 204, 124e132. Lin, Y., Zhao, Y.F., Qiu, X.H., Ma, J., Yang, Q.Y., Shao, M., Zhu, T., 2013. Spatial distri- Zhao, X.F., Zhang, H.J., Fan, J.F., Guan, D.M., Zhao, H.J., Ni, Y.W., Li, Y., Chen, J.P., 2011. bution of polychlorinated naphthalenes in the atmosphere across North China Dioxin-like compounds in sediments from the Daliao River Estuary of Bohai based on gridded field observations. Environ. Pollut. 180, 27e33. Sea: distribution and their influencing factors. Mar. Pollut. Bull. 62, 918e925. Liu, G.R., Zheng, M.H., Du, B., Nie, Z.Q., Zhang, B., Liu, W.B., Li, C., Hu, J.C., 2012.