Detection of Polycyclic Aromatic Hydrocarbons in Water Samples by Annular Platform-Supported Ionic Liquid-Based Headspace Liquid-Phase Microextraction

In this paper, a new method of annular platform-supported headspace liquid-phase microextraction (LPME) was designed using ionic liquid as an extraction solvent, wherein extraction stability and efficiency were improved by adding an annular platform inside the extraction bottle. The ionic liquid 1-silicyl-3-benzylimidazolehexafluorophosphate was first synthesized and proved to be an excellent extraction solvent. Coupled with liquid chromatography, the proposed method was employed to analysis of polycyclic aromatic hydrocarbons (PAHs) in water and optimized in aspects of extraction temperature, extraction solvent volume, extraction time, pH, stirring rate, and salt effect of solution. The results indicated that this method showed good linearity (R2 > 0.995) within 0.5 µg·L−1 to 1000 µg·L−1 for PAHs. The method was more suitable for extraction of volatile PAHs, with recoveries from 65.0% to 102% and quantification limits from 0.01 to 0.05 µg·L−1. It has been successfully applied for detection of PAHs in seawater samples.


Introduction
PAHs have been included in the environmental control legislation by many agencies, such as the US Environmental Protection Agency (EPA) [1] and the European Union. e maximum admissible concentration of naphthalene (1200 ng·L −1 ), anthracene (100 ng·L −1 ), fluoranthene (100 ng·L −1 ), and benzo[a]pyrene (100 ng·L −1 ) for superficial water was restricted by the European Union [2,3]. To better evaluate the impact of PAHs on ecosystems, an efficient and sensitive analysis method must be developed. However, PAHs are found at very low levels (ng·L −1 or pg·L −1 ) in environment samples. It is necessary to perform an enrichment step in sample preparation due to the lack of sensitivity obtained with conventional analytical instrumentation.
According to extraction patterns, application of ILs in the LPME technology can be classified into the following: dispersive liquid-liquid microextraction (DLLME) [25], hollow-fiber LPME [26], and suspended single-drop microextraction (SDME). Dispersive LPME requires large extraction amount with limited extraction efficiency; hollow-fiber LPME features low recovery with poor reproducibility; SDME has drawn much concern owing to its high enrichment efficiency. However, research studies showed that the sample solution should be stirred continuously during SDME, so extraction droplets easily fall off, affecting extraction stability and reproducibility. Furthermore, the droplets' volume of SDME was restricted, although adding a silicon rubber tube [12], a polytetrafluoroethylene sleeve [15], or a plastic tube [27] on a syringe needle seriously limits the improvement of extraction efficiency. erefore, the SDME device and pattern should be optimized to promote the development of SDME technology. Furthermore, ILs have played crucial roles in the aspect.
1-Octyl-3-methylimidazolehexafluorophosphate ([C 8 MIM]PF 6 ) was first used as an SDME solvent coupled with liquid chromatography to extract PAHs and aniline compounds in water samples [12]. Since then, some other types of ILs were used as an extraction solvent for SDME [28].
In this study, an IL-based headspace LPME device was designed using the homemade IL 1-silicyl-3benzylimidazolehexafluorophosphate ([SBnIM]PF 6 ) as an extraction solvent. e device was optimized by adding an annular platform inside the extraction bottle; then, IL was slowly pressed dropwise down to the annular platform by a syringe. e developed method was named "annular platform-supported IL-based headspace LPME" (APS-IL-LPME). e method altered the suspension state of extracted droplets, overcame the influence of droplet gravity, and significantly improved extraction stability and efficiency. Meanwhile, the large extraction platform can flexibly regulate addition of extraction solvent to control extraction capacity and enrichment efficiency. Considering that fluoranthene and naphthalene showed larger partition coefficients in the system of [C n MIM][PF 6 ]/water [9,29], liquid chromatography was combined with APS-IL-LPME to determine PAHs in water samples. Simultaneously, the IL [SBnIM]PF 6 containing the benzyl group in the structure of cation was used as an extraction solvent, which will be more suitable for extraction of analytes including benzene rings. e influence of different parameters on extraction was investigated and optimized in aspects of extraction temperature, extraction solvent volume, extraction time, pH, stirring rate, and salt effect of solution. Moreover, the proposed method was compared with other preconcentration methods for determination of PAHs in water samples.

Apparatus and Reagents.
A 2695 high-performance liquid chromatograph was equipped with an automatic sampler and a 2475 fluorescence detector (Waters Corp., US). An XW-80A vortex mixer (Shanghai Medical University Instrument Factory), a Milli-Q ultrapure water meter (Millipore Corp., US), and a Sorvall Biofuge Primo centrifuge ( ermo Fisher Scientific Corp., US) were used.

e Device of APS-IL-LPME.
e APS-IL-LPME device was self-manufactured in the lab. As shown in Figure 1, the whole device includes a thermostatic water bath ((2) in Figure 1), a heatable magnetic stirrer ((3) in Figure 1), and an extraction bottle with a sealing bottle cap ((1) in Figure 1). e thermostatic water bath is laid on the heatable magnetic stirrer, and the extraction bottle (made of glass) with a sealing bottle cap (made of polytetrafluoroethylene) is placed in the thermostatic water bath and fixed on an iron stand. e extraction bottle mainly consisted of the upper bottle body (ID 4.0 cm, height 4.5 cm; (5) in Figure 1), annular extraction platform (width 0.75 cm; (7) in Figure 1), and lower bottle body (ID 2.5 cm, height 4.0 cm; (6) in Figure 1), all of which are integrated as a whole from top to bottom. Inner diameter of the upper bottle body is larger than that of the lower bottle body, and the annular extraction platform is a platform with high inner and outer sides and a sunken middle part, which is used to bear the weight of extracted droplets.

Preparation of ILs.
e IL [SBnIM]PF 6 was prepared as shown in Figure 2. Firstly, 1-benzylimidazolium was synthesized [30]. en, [SBnIM]Cl (A) was prepared through a reaction between c-chloropropyl-trimethylsilane and benzylimidazolium, which was according to the references [31,32]. Lastly, the chloride anion was exchanged for the hexafluorophosphate anion (PF 6 − ) [33]. e general procedures for the synthesis of 1-benzylimidazole were as follows: formaldehyde water solution (36%, 8.35 g) and glyoxal water solution (32%, 18.1 g) were added to a 250 mL, three-necked flask provided with a stirrer and a reflux condenser. While the mixture was heated to 50°C, a mixture of benzylamine solution (10.7 g in 50 mL methanol) and ammonia solution (28%, 6.05 g) was added dropwise. After the mixture refluxed for an additional 4 h at 50°C, the methanol and water were removed under reduced pressure. e residue was then concentrated and dried in a vacuum desiccator at 60°C. e product of 1-benzylimidazole was obtained with the yield about 75%.
Secondly, 1-benzylimidazole (0.05 mol, 4.60 g) was added dropwise to a solution of c-chloropropyltrimethylsilane (0.05 mol, 17.3 g in 25 mL toluene). e solution was then refluxed for 72 h under N 2 atmosphere. e resultant was washed with deionized water to remove the residue of reactants. After concentrating and drying, the light-yellow transparent and viscous compound [SBnIM]Cl (A) was obtained and the yield was about 90%.
en, the IL [SBnIM]PF 6 (B) was prepared through anion-exchange reactions. e process was as follows: the compound [SBnIM]Cl (A) was dissolved in water, and then 1 M equivalent of KPF 6 in water was added dropwise. After stirring for 24 h at room temperature, the resultant was dissolved in dichloromethane and washed with deionized water. e silver nitrate test was used to confirm absence of chloride. Finally, the product was concentrated and dried, forming a light-yellow viscous compound [SBnIM]PF 6 (B), and the yield reached 90%. e final product [SBnIM]PF 6 was confirmed by proton nuclear magnetic resonance ( 1 HNMR) spectroscopy and Fourier transform infrared (FT-IR) spectroscopy.    Journal of Analytical Methods in Chemistry 2.4. Extraction Procedures. First, 10 mL of the PAH-spiked water sample (5.0 µg·L −1 ) was transferred to the extraction bottle, and sodium chloride was added to saturated concentration. e solution was inside the lower bottle body and remained below the annular extraction platform, which was headspace LPME. e sealing bottle cap was opened; the front end of the microsyringe (100 µL) was placed in the middle position of the annular extraction platform (as shown in Figure 1(b)). en, the syringe was slowly pressed down to release ILs (50 µL) on the annular platform. e sealing bottle cap was fastened down; extraction was then implemented through stirring with the required water bath temperature. After extraction, the sealing bottle cap was loosened, the microsyringe was directly used to absorb IL on the annular platform, the solution was diluted to 100 µL using acetonitrile, and 10 µL of the solution was absorbed through the automatic sampler for high-performance liquid chromatography (HPLC) analysis.

Analysis Conditions.
e analytes were separated on a Waters PAHC 18 chromatographic column (250 mm × 4.6 mm, ID 5 µm, Waters Corp.) at a column temperature of 40°C. e sample injection volume was 10 µL. e mobile phase consisted of acetonitrile (A) and water (B). e total run time was 20.0 min with a flow rate of 1.0 mL min −1 .
e gradient elution steps were as follows: 0-10.0 min, A increased from 60% to 100%, then remained at 100% until 16.0 min, and thereafter returned to 60% in 1.0 min, followed by reequilibration time for 3.0 min. e parameters of emission wavelength (λ em ) and excitation wavelength (λ ex ) for PAHs on a fluorescence detector are shown in Table 1.

Optimization of Extraction Conditions.
Extraction efficiency of ILs for the analytes was mainly influenced by several parameters. In this study, optimized parameters included volume of the extraction solvent, extraction temperature, extraction time, pH, stirring rate, and salt effect of solution. During optimization of parameters, the PAHspiked water sample (5.0 µg·L −1 ) was used as an extraction substrate. Extraction efficiencies of methods were evaluated through enrichment factors (EFs) and recoveries. Computational formulas were expressed as follows: where C t and n t refer to the concentration and molar quantity of PAHs in the extraction phase, respectively; C 0 is the initial concentration of PAHs in the water sample before extraction; V t corresponds to the final volume of the extraction phase; and V 0 stands for the volume of the water sample before extraction.

Volume of ILs.
Volume of the extraction solvent is an important factor influencing extraction efficiency. is study assessed influences of different volumes (10, 20, 50, 80, and 100 µL) of the IL [SBnIM]PF 6 on extraction efficiency of six kinds of PAHs ( Figure 3). When volume of IL was within 10 µL to 50 µL, extraction efficiency gradually improved as volume enlarged; enrichment efficiency remained within 50 µL to 80 µL but decreased when IL volume increased to 100 µL. us, as volume of the extraction solvent enlarged, extraction amount (n t in Formula (1)) of analytes correspondingly increased. However, when volume (V t ) of the extraction solvent escalated too rapidly, analyte concentration (C t ) and EFs decreased. erefore, to obtain maximum enrichment efficiency and good reproducibility, volume of the IL extraction solvent was set at 50 µL.

Extraction Temperature.
To obtain satisfactory enrichment efficiency, the added water sample should not be higher than the level of the annular extraction platform. e extraction process involved headspace LPME. Extraction temperature directly influenced velocity and concentration of analytes released from the water sample and then influenced extraction efficiency indirectly. e increase in extraction temperature will shorten equilibrium time, accelerate analysis, and increase concentration of analytes in the gas phase. However, extraction for analytes was an absorption process. us, high temperature will reduce the distribution coefficient of analytes in the extraction solvent and also diminish extraction amount. erefore, this study also assessed influences of different temperatures on extraction efficiency (Figure 4). When the temperature increased from room temperature to 50°C, EFs of six kinds of PAHs increased rapidly but slowed down under 50°C to 60°C and decreased under 70°C. us, in this study, 50°C was selected as an optimal extraction temperature.

Extraction Time.
eoretically, before reaching the equilibrium during extraction, prolonged duration will increase contents of analytes absorbed by the extraction solvent. is study also evaluated the influence of extraction time on extraction efficiency under 50°C ( Figure 5). EFs of different kinds of PAHs improved rapidly before 30 min and improved slowly afterward. In this study, the extraction time was set to 30 min to guarantee EFs of analytes. However, according to Figure 5, if larger EFs are desired, an extraction time up to 50 min would be adopted.

pH of Solution.
To study the influence of pH on extraction efficiency, 1 mol·L −1 HCl or NaOH was used to regulate pH of the extracted water sample within 2 to 13. e eoretically, before reaching the equilibrium, accelerating the stirring rate will increase contents of analytes absorbed by the extraction solvent. e results showed that extraction efficiency of all analytes slightly increased as the increase of stirring rate was up to 1000 rpm, while the reproducibility was poor after the stirring rate greater than 1000 rpm. A stirring rate of 1000 rpm was selected in this work.

Salt Effect.
In headspace LPME, adding NaCl in the sample solution can reduce solubility of analytes in the water phase, improve the gas-liquid partition coefficient, and enlarge concentration of analytes in the gas phase to increase extraction amount indirectly. However, extremely high NaCl concentration will hinder transport of analytes in the sample solution and influence their diffusion in the gas phase. is study assessed the effects of NaCl addition (0%-36%, w/v) on extraction efficiency ( Figure 6). e results showed that enrichment efficiency was the highest when the NaCl solution was saturated, indicating that reducing solubility of PAHs in the water phase exerted a significant influence. erefore, the optimum concentration for salt addition was saturated NaCl solution.

Method Evaluation.
To evaluate practical applicability of the APS-IL-LPME method, reproducibility, limit of detection (LOD), and limit of quantification (LOQ) were investigated with PAH-spiked ultrapure water under the optimum conditions, the results of which are summarized in Table 2. It showed that the method presented good linearity (R 2 > 0.995) within 0.5 µg·L −1 to 1000 µg·L −1 for the six kinds of PAHs. Replicate extractions of spiked ultrapure water solution (5.0 µg·L −1 ) were carried out on intraday and interday to show precision of the method. e relative standard deviation (RSD) of PAHs ranged from 4.7% to 11.6% in all cases, and LOQ was within 0.01 µg·L −1 to 0.05 µg·L −1 . e recoveries of Nap, Acp, Flu, Ant, and FL were within 70.2% to 102%, whereas only that of Pyr was lower than 70%. Meanwhile, EFs ranged from 65.0 to 102, which were comparable with those reported by Liu et al. [12]. e comparison of different methods for determination of PAHs in water is shown in Table 3. LLE [34] used less extraction time and exhibited low LOQ, but it was relatively solvent-consuming and had lower extraction efficiency. e SPE [35] method gave higher extraction efficiency. However, it exhibited high LOQ and needed more sample volume, while the operation of SPME [36] was complex and timeconsuming. e LPME combined with gas chromatography [25,37,38] showed higher LOQs. e IL-SDME method [12] showed high EFs, while had lower extraction efficiency (<50%), calculated according to Formula (2) in Section 3.1.
e previous reported IL-LPME methods [12,22,39,40] all used [C 8 MIM]PF 6 as an extraction solvent, and the LODs of most of them were higher than those of the proposed method. In conclusion, the proposed APS-IL-LPME method is simple, stable, and an advantageous alternative, the extraction efficiency and LOQ of which are better than or comparable with those of other methods.

Application to Real Water Samples.
Seawater samples collected from Shazikou in Qingdao and corresponding PAH-spiked seawater samples were extracted and detected according to the optimization method. As shown in Figure 7(c), no PAHs were found in the seawater sample. To evaluate this method, the seawater sample was spiked with PAHs at two levels (5.0 and 10 µg·L −1 ). As can be seen in Table 4, the recoveries for six PAHs were from 67.2% to 106% with the RSD between 3.5% and 10.4%. Compared with the recoveries of PAHs in spiked ultrapure water solution, the effects of matrix constituents were minor.   Table 4, EFs and recoveries decreased gradually with increasing boiling point and molecular weight due to decreased molecule diffusion and gas-liquid partition coefficient.
e results were consistent with the   findings of headspace LPME reported by other studies [12], indicating that this method is more suitable for enrichment of volatile PAHs in water samples due to the nonvolatile property of IL.

Conclusions
is work proposed a new method of APS-IL-LPME using the IL [SBnIM]PF 6 as an extraction solvent, which overcame the influence of IL gravity and significantly improved extraction stability and efficiency. e IL [SBnIM]PF 6 was homemade and proved to be an excellent extraction solvent for LPME. Combined with liquid chromatography, the method was employed to detect PAHs in the water sample. e optimized conditions for extraction were under 50°C for 30 min with 50 µL extraction solvent and saturated NaCl concentration in the sample solution. e results indicated that the extraction efficiency and LOQ of the method were better than or comparable with the previously reported ones and suitable for detection of PAHs, such as Nap, Acp, Flu, Ant, FL, and Pyr, in water solutions, with EFs ranging from 65.0 to 102 and LOQ from 0.01 to 0.05 µg·L −1 . e APS-IL-LPME method was first used for enrichment of PAHs in real seawater samples, indicating that it will have broad potential in enrichment and analysis of organic pollutants in the environment and food samples. is method is especially advantageous for the enrichment of volatile compounds and shows higher extraction stability than suspended SDME by adding an annular platform. However, as the headspace LPME mode, the amount of water samples is limited and not higher than the level of annular platform. Moreover, in this work, nonvolatile compounds are not transferred to the headspace and not extracted by the ionic liquids. In the future work, we will continue to develop a new extraction mode to apply this method in immersed LPME for extraction of nonvolatile compounds.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no financial and commercial conflicts of interest in connection with the work submitted.