Journal 
INTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs) are organic compounds with multiple fused aromatic rings (two to seven rings) generally released by various human activities including the burning of fossil fuels, the incomplete burning of carbon-containing materials (e.g. gasoline and coal), tire particles, motor oil spills, vehicle exhaust, and crumbling asphalt (WHO 2000). Therefore, the contamination of PAHs in the agricultural waterways in Korea can be predictable since many of them are located between car roads and agricultural lands. PAHs do not biologically break down due to their aromatic ring structures and some PAHs (e.g. benzo(a) pyrene) are known to be carcinogenic and mutagenic (IARC 2010). In addition, noncarcinogenic PAHs including phenathrene, fluoranthene and pyrene in water can react with nitrite (NO2-) in the air and form nitrated PAHs (nitro-PAHs) known to be more toxic than the parent PAHs (Shailaja et al. 2006). It should be noted that the water from the waterways is mainly used by Korean farmers to supply their crop and increasing vehicular traffic or air pollution may also cause increase in levels in PAHs around the waterway environment (Lin et al. 2015).
To prevent the serious contamination, it is important to continuously monitor the levels of PAHs in the water environment using a more simplified and advanced analytical method. In many studies on monitoring and assessment of PAHs, conventional extraction methods combined with gas chromatography mass spectrometry (GC-MS) have been adopted to determine the concentration of PAHs in various environmental samples (Maliszewska-Kordybach et al. 2008; Huang et al. 2014). In this study, QuEChERS (quick, easy, cheap, effective, rugged and safe) extraction method combined with GC tandem mass spectrometry (GC-MS/MS) was modified for PAHs analysis. The QuEChERS method is effective but greatly reduces analysis time and costs much less than the traditional extraction method involved with the large amount of organic solvent, tedious concentration procedures or cross contamination. This method was originally introduced for the determination of pesticide residues in agricultural produce (Anastassiades et al. 2003). It has been gradually extended to the analysis of organic contaminants such as mycotoxins (Jettanajit et al. 2016), veterinary drugs (Guo et al. 2015) and PAHs. Pesticide residues and PAHs in fresh herbs were determined using a modified QuEChERS method and GC-MS (Sadowska-Rociek et al. 2014).
GC-MS/MS were mainly used for multiresidue pesticide analysis (Zou et al. 2016) but it became the analytical tool of choice when target compounds are analyzed in a very complicated matrix (e.g. environmental samples and drug residues or metabolites in biological samples). The target analytes can be detected regardless of the sample matrix or other coeluted compounds of interest (or interferences) since the selected precursor or parent ions from the first stage of MS/MS are induced to further dissociate by collision energy. In this manner, MS/MS can confirm the compounds at trace levels (ppb level) lower than single MS can detect (Hayward et al. 2015). Multiclass pesticides and PAHs in fatty fish were analyzed using GC-MS/MS at trace level and recoveries at 10, 25 and 50 μg kg-1 ranged from 60 to 120% with RSD<11% (Chatterjee et al. 2016).
This study was aimed to optimize modified QuEChERS extraction and dispersive solid phase extraction (d-SPE) clean-up methods to extract and purify PAHs in water samples for the GC-MS/MS analysis. The optimized analytical method was applied to evaluate the level of PAHs in the waterways in Chungbuk (15 sites) and Gyeongbuk (6 sites). These results were also assessed associated with possibles PAHs sources. The determined PAHs were also assessed their environmental risks using ECOSAR. With these results, safety in agricultural ecosystems will be kept from persistent organic pollutants including PAHs.
MATERIALS AND METHODS
1. Chemicals
Certified reference standards of each PAH had >98% purity and were purchased from Sigma Aldrich (US). Acetonitrile (ACN) was purchased from Merk (Darmstadt, Germany). QuEChERS salt packets (4 g anhydrous MgSO4, 1g NaCl, 1g trisodium citrate dehydrate and 0.5 g disodium hydrogen citrate sesquihydrate) were purchased from Ultra Scientific (Seoul, Korea) and d-SPE tubes containing primary secondary amine (PSA, 25 mg) and anhydrous magnesium sulphate (MgSO4, 150 mg) were purchased from Restek® (US).
2. Sample Preparations
ACN (10 mL) was added to water sample (10 mL) in 50 mL of poly propylene plastic tube and vigorously shaken for 1 min. While contacting crushed ice with an extraction tube in order to cool down the extract, the salt was added to the tube for the partitioning and vigorously shaken for 1 min and centrifuged for 5 min (3,500 rpm). For clean-up, 1 mL of each extract was added to 2 mL of d-SPE tube containing PSA and MgSO4 and vigorously shaken for 1 min. After being centrifuged (15,000 rpm), 0.5 mL of extracts were taken into a GC vial and evaporated under the gentle stream of nitrogen gas. ACN (0.1 mL) was added to the vial for the GC-MS/MS analysis.
3. Instrumental Analysis
The sample analysis was carried out with Bruker SCION TQ (Triple Quadrupole mass spectrometry) in SRM (selected reaction monitoring) with CP-8400 autosampler. For separation of the analytes, ZB-SemiVolatiles (30 m×0.25 mm×0.25 μm film thickness, Phenomenex) was used. The injection volume was 2 μL with splitless pulsed pressure mode at 250 kPa. The carrier gas was helium with constant flow (1.0 mL min-1) and inlet temperature was set at 300℃. The initial oven temperature with 90℃ was held for 3 min, ramped into 170℃ at 25℃ min-1 and ramped into 300℃ at 5℃ min-1 (held for 4 min). Mass spectrometer was run with EI (electron-impact ionization) mode with 70 eV and transfer line and ion-source temperatures were set at 300 and 280℃, respectively.
4. Method Validation and Fortification Studies
Linearity, limits of quantification (LOQs), method detection limit (MDL), limits of detection (LODs), accuracy and precision were determined during validation of the analytical method. External standard method was employed for quantitative analysis. For accuracy and precision, the recovery and reproducibility experiments were performed by fortifying blank agricultural water sample (10 mL) in three replicates at two spiking levels: 10 and 50 μg L-1. Using matrix matched standards, the quantification of each compound was carried out by plotting a calibration curve (1, 5, 10, 25, 50, 100, 250, and 500 μg L-1) for linearity (R2>0.99). LOQs of the optimized method were calculated as the lowest concentration giving signal-to noise ratios of 10 (S/N=10) by injecting matrix matched calibration standard into GC-MS/MS. MDL was calculated by multiply LOQs by five, the concentration factor obtained in sample preparation.
5. Agricultural Water Sample Analysis
Water samples were obtained from Chungbuk (15 sites) (site 1: 37°02ʹ25.9ʺN 127°55ʹ31.5ʺE, 2: 37°02ʹ27ʺN 127°55ʹ1ʺE, 3: 37°02ʹ29ʺN 127°54ʹ45.0ʺE, 4: 37°02ʹ32.2ʺN 127°54ʹ34.9ʺE, 5: 37°02ʹ38.5ʺN 127°54ʹ20.0ʺE, 6: 37°02ʹ52.5ʺN 127°53ʹ29.9ʺE, 7: 37°02ʹ22.5ʺN 127°55ʹ46.8ʺE, 8: 37°02ʹ14.6ʺN 127°56ʹ12.2ʺE, 9: 37°02ʹ13.3ʺN 127°56ʹ46.0ʺE, 10: 37°02ʹ20.4ʺN 127°57ʹ03.2ʺE, 11: 37°02ʹ45.4ʺN 127°57ʹ27.1ʺE, 12: 37°05ʹ03.3ʺN 127°57ʹ22.5ʺE, 13: 37°05ʹ21.1ʺN 127°57ʹ35.0ʺE, 14: 37°04ʹ28.6ʺN 127°55ʹ50.3ʺE, 15: 37°02ʹ25.9ʺN 127°55ʹ31.5ʺE) and Gyeongbuk (6 sites) (1: 35°37ʹ44.9ʺN 128°26ʹ06.1ʺE, 2: 35°57ʹ19.6ʺN 128°20ʹ05.6ʺE, 3: 35°57ʹ00.7ʺN 128°19ʹ37.7ʺE, 4: 35°55ʹ57.1ʺN 128°34ʹ02.3ʺE, 5: 35°35ʹ58.4ʺN 128°28ʹ28.4ʺE, 6: 35°53ʹ07.3ʺN 128°36ʹ04.9ʺE). Replicate samples from Gyeongbuk were collected. All the samples were prepared using the analytical methods proposed in this study. For the quantitative analysis of water samples, matrix matched standards were prepared with blank agricultural water sample, pre-checked by the preliminary analysis.
Measured PAHs were evaluated for their environmental risks using risk characterization ratios (RCRs) in relation to predicted no-effect concentrations (PNEC) calculated by ECOSAR (version 2.0). If RCR value is greater than 1, risk is not controlled. Otherwise, risk is adequately controlled if the RCR value is less than 1.
RESULTS AND DISCUSSION
1. GC-MS/MS Analysis and Sample Preparation
For the optimization of the MS parameters, all the analytes were initially monitored in full scan mode in the 50- 550 m/z ranges and then, one or two precursor ions (or parent ions) for each analyte were selected (by considering selectivity (specificity) and sensitivity. When the precursors (or parent ions) were dissociated once more by the collision energy, the best product ions were selected for each transition in a multiple reaction monitoring (MRM) experiment. The most intense transition of each analyte was selected for the quantifier analysis and the second most intense for the qualifier analysis. Due to structural stability of aromatic rings (two to seven rings) in PAHs, high collision energies were required to produce product ions. The quantification and identification ions and collision energy for each PAH selected in MRM experiment are shown in Table 1. The consequential specific product ion spectrum confirmed the target compound identified by retention time of each compound. In this way, selectivity of MS/MS enhanced the signal to noise so low limit of detections (LODs) for the PAHs (e.g. LOD of phenanthrene is 1 μg L-1) were achieved in this experiment. MDLs were 0.2 μg L-1 for all the target compounds. Some studies showed that LOD of phenanthrene ranged from 2.5 to 3.5 μg L-1 using GC-MS (Zhao et al. 2015; Zou et al. 2016). It was demonstrated that signalto- noise ratios of PAHs determined using GC-MS/MS was much superior to those of them using GC-MS (Zou et al. 2016).
The QuEChERS extraction method is a streamlined approach recently developed to separate organic compounds (pesticide residues or PAHs) in a complex matrix (Anastassiades et al. 2003). This technique provides an environmentally friendly alternative to conventional liquid-liquid and solid phase extractions, requiring the large amount of organic solvent. The procedure involves two simplified steps; sample extraction and d-SPE clean-up. In this study, the sample (10 mL) extraction and partitioning were carried out using a minimal amount of organic solvent (10 mL) and salt. Then, the target analytes in extract were purified using a 2 mL d-SPE tube containing PSA alone. The RSD data validated that the proposed method was reproducible. Fourteen PAHs provided recoveries in the range of 60-110% at the two fortified levels of 10 and 50 μg L-1 associated with RSDs (<20%) (Table 2). Good linearity (R2>0.99) for all the target analytes was achieved for both solvent and matrix matched standards for quantification. Phenanthrene, anthracene and 1-acenaphthenone gave lower recoveries (approx. 70%) rates than other compounds (85-110%). These compounds must be lost during the evaporating procedure since they have lower molecular weight and more volatile than other PAHs (Vane et al. 2014).
2. Levels of PAHs in Water Samples
To determine the concentration of 14 PAHs in real water samples, water was sampled in the waterways located in Chungbuk (15 sites) and Gyeongbuk (6 sites), Korea. Except phenanthrene and fluoranthene, all the PAHs analyzed in the water samples were not detected. Phenanthrene was detected in all the water samples from both Chungbuk and Gyeongbuk areas. The concentration of phenanthrene ranged from 0.54 to 2.53 μg L-1 and the distribution of phenanthrene and fluoranthene in both places is shown in Fig. 1A and B. One of the sampling sites in Gyeongbuk gave the highest concentration (2.53 μg L-1) of phenanthrene out of other sites (Table 3). Considering that the sampling area is located near a tire factory, the compound was possibly derived from runoff of oil spills or tire particles. It may be one of petrogenic PAHs, exclusively dominated by the C1 to C4 alkylated homologues of certain parent PAHs, in particular, naphthalene, phenanthrene, dibenzothiophene, fluorene, and chrysene (Wang et al. 2007).
On the other hand, fluoranthene was detected in water samples from only two sites (0.11 and 0.16 μg L-1, respectively) in Gyeongbuk (Fig. 1B). Considering that the presence of fluoranthene indicates less efficient or lower-temperature combustion, this compound may be possibly due to combustion of the vehicles passing by the waterways.
Considering both compounds in waterways are less toxic than other PAHs and “not classifiable as to its carcinogenicity to humans” (IARC 2010), water from waterways from two areas may be safe for crop irrigation. However, it should be noted that these compounds can react with NO2- in the atmosphere, producing nitro PAHs such as nitro-phenanthrene or nitro-fluoranthene. Lin et al. demonstrated that the presence of NO2- would cause the PAHsnitro- PAHs transformation. Considering that nitro-PAHs tended to be more toxic, the combined pollution of PAHs and NO2- may threaten the ecosystem. Even though PAHs are not detected in the most water samples, this monitoring should be expanded to analyze other toxic compounds (e.g. POPs) or levels of PAHs in sediment should be determined. Zhang et al. (2004) demonstrated that the concentration of total PAHs in sediment was higher than those in surface water. According to the study, the PAHs were dominated by 2-3-ring compounds in water samples and by 3-4-ring compounds in sediment.
With Tables 3 and 4, risks by phenanthrene in the two provinces in Korea well controlled because RCR values is less than 1 or close to the 1. The site Gyeongbuk 4 is the site, which we should monitor annually because RCR values for the three aquatic organisms is ranged from 1.77 to 3.21. Similarly, two sites of Gyeongbuk provinces (sites 2 and 4) contained fluoranthene in the waterways, however, their RCR values were below 1. It means that all sites were controlled properly for the fluoranthene.
