Liquid phase microextraction based on the solidification of a floating ionic liquid combined with high-performance liquid chromatography for the preconcentration of phthalate esters in environmental waters and in bottled beverages

Abstract

A novel liquid phase microextraction method based on the solidification of floated ionic liquids (SFIL-LPME) was developed to determine four phthalate esters (diethyl phthalate, diallyl phthalate, benzyl butyl phthalate, and dicyclohexyl phthalate) in environmental waters and bottle beverages and subsequently separate them using high-performance liquid chromatography. In this work, we customized a tributyldodecylphosphonium tetrafluoroborate ([P_{4 4 4 12}][BF₄]) ionic liquid (IL) to obtain a low density such that it can be solidified at a low temperature, and then used as an extraction solvent in liquid-phase microextraction based on the solidification of floated organic drops (SFO-LPME). This means that more potential ILs can be customized to broaden extraction solvents in SFO-LPME. To identify the significant factors that affect extraction efficiency, parameters such as the volume of [P_{4 4 4 12}][BF₄], the extraction time, the rotation speed, the extraction temperature and the ionic strength were optimized using a Plackett–Burman design. A central composite design was then used to optimize the identified significant factors to study the effect of interactions on the experiment. Under these optimized conditions, the extraction recoveries of the four phthalate esters in water samples ranged from 73.0 to 94.6% and in bottled beverage were from 70.2 to 101.3%, with relative standard deviations (RSDs) ranging from 0.6 to 8.2% and 0.6 to 7.8%. Good linearity for both water samples and bottled beverages were obtained in the range of 5–500 μg L⁻¹, with correlation coefficients greater than 0.9993. The limits of detection for the four phthalate esters varied from 0.27 to 1.1 μg L⁻¹ for water samples and 1.22 to 2.36 μg L⁻¹ for bottled beverages. The two developed methods were then successfully applied to the determination of phthalate esters in environmental waters and bottled beverages.

---

1. Introduction

Phthalate esters (PAEs) are a family of plasticizers that are used to enhance the workability, extensibility and elasticity of polymers.¹ The production of PAEs has significantly increased, and they are prevalent in many consumable and household products.² However, PAEs and their metabolites and degradation products can cause toxic effects in multiple organ systems.^3,4 Previous studies have shown that PAEs have hormone-disrupting activities that may be associated with known estrogenic and antiandrogenic activities.^5–7 Moreover, they are suspected to have carcinogenic properties.^8,9 The US Environmental Protection Agency (EPA) and several other agencies have classified PAEs as priority pollutants.^10,11

People come into contact with PAEs in daily life through a wide variety of plastic packaging materials.^12,13 Among these, bottled beverages provide direct contact because the beverages provide a direct route for PAEs to enter the human body.¹⁴ Tea, herbal tea and juice is the most widely consumed bottled beverages in china due to its popular taste and its prevention of several forms of disease.¹⁵ In addition, plastic products are significant pollution sources for water.¹⁶ Therefore, it is necessary to develop a reliable and highly sensitive method for determining of PAEs in bottled beverages and environmental waters.

Due to the low concentration of PAEs in tea beverages and environmental waters, the preconcentration of the analytes prior to analysis is essential. Liquid phase microextraction (LPME) is a common aqueous sample pretreatment method for many analytical applications because of the advantage of simple, efficient and inexpensive; it also provides higher preconcentration factors and requires lower amounts of toxic extraction agents.¹⁷ Based on these advantages, several LPME alternative pretreatment techniques have been successfully introduced, including dispersive liquid–liquid microextraction (DLLME),^18,19 hollow-fiber-based liquid-phase microextraction (HF-LPME),^20,21 single-drop microextraction (SDME)^22,23 and liquid-phase microextraction based on the solidification of floating organic drops (SFO-LPME).^24,25

SFO-LPME was first proposed by Zanjani et al. in 2007.²⁴ The key component of this method is the extraction solvent, which must have a melting point near room temperature and a low density to allow it to float on the surface of the aqueous solution. The initial stage of the extraction process involves stirring the sample for the required period of time. After extraction, the sample vial is transferred to an ice bath, and the solidified floating organic drop is then transferred to a small conical vial. After melting into a solution, a few drops are subjected to analysis.^24,26 The solidification of a floating organic drop facilitates the collection of the extraction phase. However, few extraction solvents meet these requirements;¹⁷ only 1-dodecanol, 2-dodecanol, 1-undecanol and hexadecane have been reported to possess those properties.^27–30 Moreover, due to these solvents' low polarities, certain polar analytes are difficult to extract.³¹ Therefore, a more suitable extraction solvent is necessary.

ILs, known as “designable green solvents,” possess several unique physicochemical properties, such as variable viscosity, high thermal stability and negligible vapor pressure.^32,33 Because of these advantageous properties, ILs have been widely used as extraction solvents to enrich various analytes,^34–38 e.g., extraction of aluminum by IL-DLLME,³⁹ extraction of musk fragrances by IL-SDME,⁴⁰ and extraction of alkaline compounds by IL-HF-LPME.⁴¹ IL usage has been proven easy to execute and environmentally friendly while providing high enrichment and recovery. Because ILs consist of organic cations and organic or inorganic anions, they can be structurally customized based on diverse combinations of cations and anions.⁴² ILs can be structurally tailored to be hydrophobic or hydrophilic³¹ and to possess a suitable melting point and density for use in SFO-LPME. According to the existing literature researches, some quaternary phosphonium salt and quaternary ammonium salt ILs can be customized to possess the same properties as the organic drops used in SFO-LPME.^43–46 In this work, we customize an ionic liquid ([P_{4 4 4 12}][BF₄]) with a melting point of 25 °C, that can solidify at a low temperature and has a low density at 0.97 g cm⁻³, thus allowing it to float on the surface of an aqueous solution.⁴⁷ This IL was used in the place of an organic solvent as the extraction solvent in LPME to broaden the application of the solidification of the floating organic drops method.

In this study, [P_{4 4 4 12}][BF₄] was initially used as the extraction solvent in the SFIL-LPME method. This novel method coupled to HPLC was developed to determine four phthalate esters in environmental waters and bottled beverages. The key parameters that affected the extraction efficiency, including the volume of [P_{4 4 4 12}][BF₄], the extraction time, the rotation speed, the extraction temperature and the ionic strength, were optimized by a Plackett–Burman design. The interaction effects between the important parameters were studied by means of a central composite design (CCD).

2. Materials and methods

2.1. Chemicals and samples

All of the phthalate esters (diethyl phthalate/DEP, diallyl phthalate/DAP, benzyl butyl phthalate/BBP and dicyclohexyl phthalate/DCHP), as well as sodium tetrafluoroborate (NaBF₄) were obtained from the Aladdin Reagent Corporation (Shanghai, China). Tributyldodecylphosphonium bromine ([P_{4 4 4 12}]Br) was purchased from the Center for Green Chemistry and Catalysis, LICP, CAS (Lanzhou, China). Sodium chloride (analytical grade) was supplied by the Beijing Chemical Reagent Company (Beijing, China). Deionized water was purified using a Milli-Q SP Reagent Water System (Millipore, Bedford, MA, USA). HPLC-grade acetonitrile was purchased from Dikma Limited (Beijing, China), and Cleanert PSA was supplied by Bonna-Agela Technologies Inc. (Tianjin, China).

Standard stock solutions of 100 mg L⁻¹ of each PAE were prepared in acetonitrile. The working standard aqueous solutions were prepared freshly by diluting an appropriate amount of the mixed standard solution to different concentrations in HPLC-grade acetonitrile. All solutions were stored at 4 °C.

2.2. Synthesis of [P_{4 4 4 12}][BF₄]

[P_{4 4 4 12}][BF₄] was prepared through anion exchange from the corresponding bromide salt of the phosphonium cation ([P_{4 4 4 12}]Br) with one equivalent of sodium tetrafluoroborate (NaBF₄) in deionized water. The lower density hydrophobic IL phase [P_{4 4 4 12}][BF₄] was removed by a straw and washed with water 6 to 8 times. The product was then dried under reduced pressure at 60 °C for at least 24 h. The product was characterized by ¹H NMR and FTMS analyses, ¹H NMR chemical shifts relative to tetramethylsilane (TMS) and were as follows: δ_H (300 MHz, CDCl₃) 0.87–1.01 (m, 12H, CH₂CH₃), 1.24–1.33 (m, 16H, CH₂), 1.45–1.61 (m, 16H, CH₂), and 2.11–2.28 (m, 8H, PCH₂). FTMS (ESI+) m/z calcd for [P_{4 4 4 12}]⁺ 371.38011, found 371.38037. The mass spectrum was added to the ESI as Fig. S1.†

2.3. Instruments

A chromatographic analysis of the four phthalate esters was performed on an Agilent 1200 HPLC system (California, USA) equipped with an automated sample injector, a binary high-pressure pump, a column oven and a variable-wavelength detection (VWD) system. The separation of the analytes was performed on an Agilent Eclipse Plus C18 analytical column (5 μm, 4.6 mm × 250 mm) with Spursil C18 guard cartridges (5 μm, 2.1 mm × 10 mm, Dikma Limited). A magnetic stirrer (DAIHAN, SMHS-3, Korea) was used during the microextraction procedure. ESI-MS measurements were recorded using a Fourier Transform Ion Cyclotron Resonance mass spectrometer (Bruker, Germany) equipped with an orthogonal (ESI+) source. ¹H NMR spectra were performed using a Bruker AVANCE AV-300 NMR spectrometer (Bruker, Germany).

2.4. Determination of the target compounds using HPLC

Acetonitrile (A) and water (B) were employed as the mobile phase at a flow rate of 1 mL min⁻¹, and the column temperature was 25 °C. The gradient conditions were as follows: 0–20 min, 60–100% A; 20–22 min, 100–60% A; and 22–25 min, 60% A. The VWD wavelength was 228 nm at 0–22 min for the PAEs and then 280 nm at 22–25 min to eliminate the influence of mobile phase change.

2.5. Preparation of the real samples

Three water samples, river water (Jinzhong, Shanxi Province), lack water (Haidian, Beijing), and reservoir water (Haidian, Beijing) were used for method validation. The environmental water samples were filtered through a 0.22 μm mixed cellulose membrane and stored in the dark at 4 °C prior to use.

The tea beverages (oolong and green tea), herbal tea beverages (Jiaduobao and Wanglaoji) from two brands and grape juice beverages were purchased from local supermarkets to validate the proposed method. The bottled beverage samples were filtered through 0.22 μm micropore membranes and sufficient Cleanert PSA was used to eliminate the pigment, an organic acid that exists in bottle beverages, to reduce the matrix effect. Those samples were diluted with equal amount of water prior to use.

2.6. Extraction procedure

An 8 mL volume of samples was spiked with four PAEs and introduced into a 10 mL glass vessel. A 60 μL aliquot of [P_{4 4 4 12}][BF₄] was delivered to the surface of the solution using a microsyringe. The vessel was then sealed and maintained at 30 °C. To environmental water sample, the extraction time is 2 h at a stirring rate of 1200 rpm, while 3 h at 700 rpm for bottled beverages. Under this stirring condition, the suspended IL remained in the top-center position of the aqueous samples. After the desired extraction time, the sample vessel was transferred to an ice beaker, and the IL was solidified. The solidified IL was then transferred into a conical vial using a small spatula, where it immediately melted. The IL phase was then diluted with 30 μL of acetonitrile, and 10 μL of the mixed solution was directly injected into the HPLC system for quantification.

2.7. Data handling and processing

Construction of the experimental design matrices and analysis of the experimental data was performed using the MINITAB version 16 software (Minitab Inc., USA).

2.8. Calculation of EF and ER%

The enrichment factors (EFs) and extraction recovery (ER%) were used to evaluate the effects of the experimental conditions on the extraction efficiency, and were calculated using the following equations:

where C_IL, C_water, V_IL and V_water are the concentration of the analytes in the IL, the initial concentration of the analytes within the samples, the volume of the IL and the volume of the samples, respectively.

3. Results and discussion

3.1. Optimization of the experimental conditions

In previous studies using SFO-LPME, the parameters relevant to the extraction process have been typically optimized through single-dimensional searches.^25,48 However, single-dimensional searches cannot consider the interactions among variables and are laborious and time-consuming. In this study, we combined the P–B design and CCD to determine the optimal experimental conditions. The P–B design was used to identify the significant effects, and CCD was used to assess the interaction between these significant effects and obtain the optimum conditions.

3.1.1. P–B design. The five factors in the proposed method that could affect extraction efficiency were evaluated by the Plackett–Burman design. These five factors include the volume of [P_{4 4 4 12}][BF₄], were considered for each factor (Table 1). In the matrix of the P–B design, twelve experiments were randomly carried out to evaluate the major effects of these five factors and eliminate the effects of extraneous or nuisance variables. Each experiment was performed in triplicate, and the corresponding extraction recoveries were considered the experimental response. ANOVA was used to evaluate the effects, and a 95% confidence interval was applied to determine the statistical significance. A standardized Pareto chart was generated to visualize the effects. The vertical line in Fig. 1 indicates the minimum t-value, and variables with higher t-values were considered to be statistically significant factors. Fig. 1 shows the statistically significant effect of diallyl phthalate, which was representative of PAE behavior.

Table 1 Experimental variables and levels of the Plackett–Burman design

Variables	Level
	Low (−1)	High (+1)
(A) extraction time (h)	0.5	2.5
(B) the volume of [P4 4 4 12][BF4] (μL)	30	80
(C) rotation speed (rpm)	400	1200
(D) the temperature of sample solution (°C)	30	50
(E) ionic strength (NaCl concentration; w/v) (%)	0	4

---

Fig. 1 Pareto chart of the main effects for the Plackett–Burman design.

As shown in Fig. 1, the extraction time and the volume of [P_{4 4 4 12}][BF₄] were the most influential parameters, followed by the rotation speed. In contrast, the extraction temperature and the ionic strength exhibited no significant effect on extraction efficiency. The extraction time, the volume of [P_{4 4 4 12}][BF₄] and the rotation speed can directly affect extraction efficiency, because a long extraction time and a sufficient volume of [P_{4 4 4 12}][BF₄] can ensure that the analytes are thoroughly extracted.^25,29 A high rotation speed can result in rapid mass transfer from the sample solution to the IL, which increases the extraction efficiency and shortens the extraction time.⁴⁸ Based on the above findings, the extraction temperature was set at 30 °C and the NaCl concentration was fixed at 0%, the next optimization step considered the identified significant effects: the extraction time, the volume of [P_{4 4 4 12}][BF₄] and the rotation speed.

3.1.2. CCD design. The three significant factors (extraction time, volume of [P_{4 4 4 12}][BF₄] and rotation speed) chosen from the P–B design were optimized using CCD to obtain the best response.

Based on these responses, a quadratic model can be built by a polynomial fit in CCD, as expressed by the following equation:

y = β₀ + β₁A + β₂B + β₃C + β₁₂AB + β₁₃AC + β₂₃BC + β₁₁A² + β₂₂B² + β₃₃C²

where y is the response, β₀ is the intercept, A, B and C are the independent variables, and β₁ through β₃₃ are the coefficients of the polynomial equation. This design combined a 2^k factorial design augmented with 2k additional points (star points). To fit the quadratic polynomials, at least one central point (C) was necessary to obtain properties such as rotatability or orthogonality. The star points are located ±α from the center of the experimental domain and α was selected as 1.682 (α = ∜2^k) to establish the rotatability condition of the CCD. Based on the equation n = 2^k + 2k + C, the total number of necessary design points is 20, meaning that k and C were set at 3 and 6, respectively. Using the Minitab software, the chromatographic data for each design point were used for the multivariate regression analysis, and a mathematical model was generated. The results for each term in the model are listed in Table 2, including the regression coefficients, the Student's t distribution and the corresponding p values. A good fit to the second-order polynomial equations can be observed from the experimental data. The coefficient value of determination (R²), as explained by the model, was 0.9633 for recovery. The adjusted R² statistic was higher than the R² value, which was 0.9266, and thus indicating concurrence between the experimental data and the fitted model.

Table 2 Estimated regression coefficients and analysis of variance of the predicted model for analytes recoveries

Terms	Coefficients	t value	p
Constant	0.7328	42.797	0.000
A	0.1334	9.506	0.000
B	0.0830	7.306	0.000
C	0.0779	6.860	0.000
A^2	−0.0041	−0.280	0.786
B^2	−0.0413	−3.640	0.005
C^2	−0.0179	−1.575	0.150
AB	0.0179	1.205	0.259
AC	−0.0159	−1.073	0.311
BC	−0.0202	−1.362	0.206

---

As shown in Table 2, all of the chosen factors (extraction time, volume of [P_{4 4 4 12}][BF₄] and rotation speed) and the quadratic term of the volume of [P_{4 4 4 12}][BF₄] (B) have p-values of less than 0.05, indicating that these parameters have significant effects on extraction efficiency.⁴⁹ The interactions among A, B and C reveal no significant effect. Fig. 2 shows the response surface plots of the extraction recovery model for the three factors. In Fig. 2A to C, the rotation speed, volume of [P_{4 4 4 12}][BF₄] and extraction time were held at 800 rpm, 55 μL and 1.5 h, respectively.

Fig. 2 Response surfaces for phthalate esters using the central composite design obtained by plotting (A) extraction time vs. volume of [P_{4 4 4 12}][BF₄], (B) extraction time vs. rotation speed, and (C) volume of [P_{4 4 4 12}][BF₄] vs. rotation speed.

As shown in the Fig. 2A, the recovery increased significantly as the volume of IL increased and reach a plateau at 60 μL; the IL volume was set at 60 μL as an optimized factors. As to the extraction time and rotation speed, the recovery increased significantly as extraction time and rotation speed increased. However, an excessively high speed destabilized the droplets and created air bubbles in the extraction system which led the dispersion of IL drop.⁵⁰ The results shows that the IL can be kept with good stability in environmental water samples with rotation speed of stir bar lower than of 1200 rpm and in bottled beverages the rotation speed of stir bar was 700 rpm. According to the contour plot that obtained form CCD (Fig. S2†), at the same time considering a time limit, 2.0 h extraction time was adopted for water samples. For bottled beverages, longer extraction time is need to reach extraction equilibrium, then 3.0 h extraction time is adopt for bottle beverages. Finally, based on the results of the CCD experiments, the optimum conditions selected for SFIL-LPME are as follows: an 8.0 mL sample solution with no salt addition, 60 μL of [P_{4 4 4 12}][BF₄] used for the extraction solvent, 30 °C was adopt for extraction temperature, a rotation extraction for 2 h at 1200 rpm for water samples and 3 h at 700 rpm for bottled beverages.

3.2. Method validation

The optimum conditions obtained by the response surface model were used to perform a series of experiments to evaluate the accuracy of the SFIL-DLLME method for the PAE detection. Each concentration level was replicated three times and the pertinent results are summarized in Table 3. For water samples, good linearity were obtained in the range of 5–500 μg L⁻¹ for all four PAEs, with the R² ranging from 0.9996 to 0.9999. The recovery is between 83.5 and 102.5%, with RSDs in ranged from 2.2–5.2% and enrichment factors in the range of 120 to 164. The LODs for the PAEs, calculated at a signal-to-noise ratio of 3, were in the range of 0.27–1.14 μg L⁻¹. For bottled beverages, DAP and DEP were spiked in the range of 5–500 μg L⁻¹ and BBP and DCHP were spiked from 10 to 500 μg L⁻¹ were used to evaluate the linearity of the method. The calibration curves demonstrated good linearity within the concentration range, with R² ranging from 0.9993 to 0.9997. The RSDs was in ranged from 2.3–6.4%. The LODs for the PAEs were in the range of 1.22–2.36 μg L⁻¹. The extraction recoveries ranged from 76.9 to 91.6%, with enrichment factors in the range of 114 to 136. These results show that the two developed methods were capable of detecting PAEs in environmental water samples and bottle beverage samples.

Table 3 The performance characteristics of the SFIL-LPME method combined with HPLC-UV analysis (n = 3)

Analytes	Linearity equation	Linearity (μg L^−1)	R^2	RSD (%)	Enrichment factor	LOD (μg L^−1)	Recovery (%)
a Method validation for water samples that the extraction condition set as rotation for 2 h at 1200 rpm.b Method validation for bottled beverage samples that the extraction condition set as rotation for 3 h at 700 rpm.
DEP	Y = 122.8X + 562.3^a	5–500	0.9999	2.2	133.6	0.33	85.4
	Y = 117.1X + 121.1^b	5–500	0.9993	2.3	114	1.22	76.9
DAP	Y = 135.0X + 145.81	5–500	0.9997	2.2	156.7	0.27	100.3
	Y = 133.2X + 74.4	5–500	0.9997	2.5	136	1.46	91.6
BBP	Y = 123.3X − 100.9	5–500	0.9996	5.2	164.4	1.1	102.5
	Y = 121.5X − 3.2	10–500	0.9997	6.4	134	2.36	90.3
DCHP	Y = 91.6X − 54.4	5–500	0.9999	4.5	119.7	0.79	83.5
	Y = 90.0X + 71.4	10–500	0.9994	6.0	120	1.92	80.8

---

The proposed SFIL-LPME method is compared with other relevant methods^1,5,51,52 for extracting and determining PAEs in Table 4. It can be seen that the proposed method exhibits better linearity and extraction recovery. Moreover, the SFIL-LPME method used [P_{4 4 4 12}][BF₄] instead of a volatile organic solvent as the extraction solvent and avoided the use of a dispersant, thus making a truly “green” microextraction possible. These results reveal that SFIL-LPME is simple, effective, easy to perform, and environmentally friendly. Moreover, this method can also be used with SFIL-DLLME.

Table 4 Comparison of the proposed SFIL-LPME method with other methods for the determination of phthalate esters (n = 3)^a

Method	Extraction solvent	Organic solvent usage (μL)	Linearity (μg L^−1)	R^2	LOD (μg L^−1)	Recovery (%)	Ref.
a CAP: surfactant coacervation phase; SDBSA: sodium dodecylbenzenesulfonic acid; USAEME: slow-injection ultrasound-assisted emulsification-microextraction.
CAP-D-μ-SPE-HPLC	SDBSA	—	2–1000	0.9952–0.9992	1.0–3.8	48.6–84.8	51
DLLME-HPLC	CCl4	791	5–5000	0.9992–0.9998	0.64–1.8	84–113	52
IL-DLLME-HPLC	[C8MIM][PF6]	100	10–1000	0.9983–0.9999	1.5–4.2	54.6–93.9	1
USAEME-HPLC-UV	C2H2Cl4	40	5–300	0.9992–0.9997	0.26–1.46	76–87	5
SFIL-LPME-HPLC	[P4 4 4 12][BF4]	30	5–500	0.9996–0.9999	0.27–2.36	76.9–102.5	This work

---

3.3. Analysis of real samples

Three environmental waters and five bottled beverages (two tea beverage, one grape juice and two herb tea beverage) and were used for the determination of PAEs to study the applicability of the SFIL-LPME method. The results showed that no target PAE residues were detected in the samples by this method. Fig. 3 shows the typical chromatograms of the tea samples. The bottled beverages and water samples were spiked at two levels, 50 μg L⁻¹ and 100 μg L⁻¹, for the recovery experiment. The chromatograms of the PAEs at the two spiked levels for the green tea are displayed in Fig. 3(b) and (c), and the results are summarized in Tables 5, 6 and S1.† It can be seen that the recoveries ranged between 73.0% and 94.6% for water sample with RSD in the range of 0.8% and 6.6% for intra-day, recovery from 72.7% to 90.1% with RSD from 0.6% to 8.2% for inter-day, which show that no significant influence on the reproducibility. For bottled beverage, the recovery were between 70.2% and 101.3% and 71.2% to 91.9% for intra-day and inter-day, with RSD in the range of 0.8 to 6.6% and 0.6% to 7.8% for inter-day, it also show that good reproducibility can obtained for the determination of bottled beverage samples. These results demonstrated no significant matrix effects on the extraction, and this method is thus feasible for the determination of PAEs in water samples and bottled beverages.

Fig. 3 The HPLC chromatograms of phthalate esters in the spiked and blank green tea beverages: (1) diethyl phthalate; (2) diallyl phthalate; (3) benzyl butyl phthalate; (4) dicyclohexyl phthalate. Chromatograms a–c: spiked levels were 0, 50 and 100 μg L⁻¹, respectively.

Table 5 Spiked recoveries (%) of the three water samples (n = 3)

Analytes	Spiked level (μg L^−1)	Intra-day						Inter-day
		River water		Lack water		Reservoir water		River water		Lack water		Reservoir water
		RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)
DEP	50	79.8	1.6	76.4	4.8	79.1	4.6	76.6	3.6	76.2	7.0	75.1	3.5
	100	73.0	1.7	86.6	6.1	77.0	0.8	74.1	2.4	72.7	5.3	73.9	1.4
DAP	50	88.8	1.3	89.4	5.6	90.3	6.6	89.4	8.1	87.1	5.8	90.1	8.2
	100	88.4	4.5	94.5	4.1	87.7	3.0	89.9	2.7	84.3	5.2	88.0	2.2
BBP	50	82.0	4.1	83.1	5.7	81.8	5.5	79.0	7.4	82.7	4.0	78.9	3.8
	100	83.5	4.2	82.1	6.6	82.9	5.3	81.8	4.2	73.2	2.8	76.4	3.9
DCHP	50	78.2	6.5	86.8	3.5	77.9	3.7	76.1	4.0	75.0	5.4	84.1	7.5
	100	84.1	4.6	76.5	5.7	84.1	4.9	82.1	7.5	82.5	0.7	78.5	7.5

---

Table 6 Spiked recoveries (%) of the two tea samples (n = 3)

Analytes	Spiked level (μg L^−1)	Intra-day				Inter-day
		Oolong tea		Green tea		Oolong tea		Green tea
		RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)	RE (%)	RSD (%)
DEP	50	73.0	1.3	77.9	5.3	78.9	2.8	74.3	5.8
	100	76.8	5.9	80.2	0.9	74.5	0.9	75.4	0.6
DAP	50	91.1	3.6	97.3	4.1	91.9	4.3	87.7	6.1
	100	96.4	6.5	95.3	3.7	84.1	3.5	90.3	2.1
BBP	50	101.0	3.2	91.7	6.0	86.9	7.4	90.0	5.4
	100	101.3	4.6	94.3	3.8	84.1	5.3	83.3	1.8
DCHP	50	96.5	3.5	77.7	5.4	81.7	3.1	81.1	6.0
	100	83.7	4.6	81.4	5.0	84.5	3.6	79.2	7.8

---

3.4. Comparison of IL with organic drops as extraction solvent in DLLME-SFO

In this method, IL was initially used as the extraction solvent instead of organic drops in the SFO-DLLME. ILs are superior to organic drops as an extraction solvent in three circumstances. First, because of their designable properties, more potential ILs can be customized with low density and low temperature solidification to overcome the limitations of extraction solvents in the SFO-LPME method.^44,53 Second, ILs possess wider application possibilities in sample pretreatment⁵⁴ than do the extraction solvents that are frequently used in SFO-LPME, such as 1-undecanol and 1-dodecanol. Third, the high thermal stability and the low volatility of ILs make them safer and healthier for experiments, as they are considered environmentally benign replacements for toxin organic drops.³² However, some IL properties are not advantageous to this method. The higher viscosity of ILs can slow the mass transfer, meaning that more time is required to obtain extraction equilibrium;⁵⁵ this problem can be solved by using the SFIL-DLLME method. In addition, ILs require more time to finish solidification procedures than organic drops; this can be addressed through solidification at temperature lower than ice baths or by using ILs in low density solvent liquid phase microextraction directly, which does not require a solidification procedure.³¹ In conclusion, ILs can widely broaden the application of the SFO-DLLME method in the preconcentration of analytes.

4. Conclusions

This study successfully applied a new SFIL-LPME combined with HPLC-UV to determine four phthalate esters in environmental waters and bottled beverages. A P–B design was used to identify the significant experimental parameters, which were then optimized using CCD. The extraction solvent, [P_{4 4 4 12}][BF₄], was initially used in SFO-LPME, which is limited by the selection of an extraction solvent. The customized [P_{4 4 4 12}][BF₄] exhibited the same properties as organic drops, thus extending the range of viable extraction solvents for SFO-LPME. However, SFIL-LPME method have the drawback of long extraction time, I think this can be solved through the SFIL-DLLME method. We also believe that more ILs can be used in SFIL-LPME by using structural customization. The resulting technique provides good linearity, repeatability and high extraction recovery for each compound. Therefore, the proposed method is recommended as a simple, effective, and environmentally friendly way to perform sample preparation.

Acknowledgements

This work was supported by the fund National Natural Science Foundation of China (Project no. 21507159, 21277172 and 21377163) and Chinese Universities Scientific Fund (Project no. 2015QC074).

References

1. Y. Fan, S. Liu and Q. Xie, Talanta, 2014, 119, 291–298.
2. Y. Guo, L. Wang and K. Kannan, Arch. Environ. Contam. Toxicol., 2014, 66, 113–119.
3. F. A. Arcadi, C. Costa, C. Imperatore, A. Marchese, A. Rapisarda, M. Salemi, G. R. Trimarchi and G. Costa, Food Chem. Toxicol., 1997, 36, 963–970.
4. M. Petrović, E. Eljarrat, M. J. L. de Alda and D. Barceló, Trends Anal. Chem., 2001, 20, 637–648.
5. S. Li, T. Li, P. Gao, R. Lu, W. Zhou and H. Gao, J. Chromatogr. Sci., 2014, 52, 1127–1134.
6. Q. Xu, X. Yin, M. Wang, H. Wang, N. Zhang, Y. Shen, S. Xu, L. Zhang and Z. Gu, J. Agric. Food Chem., 2010, 58, 11311–11317.
7. L. Chang, P. Bi, X. Li and Y. Wei, Food Chem., 2015, 177, 127–133.
8. X. Cao, Compr. Rev. Food Sci. Food Saf., 2010, 9, 21–43.
9. M. Xu, M. Liu, M. Sun, K. Chen, X. Cao and Y. Hu, Talanta, 2016, 150, 125–134.
10. Y. Luo, Q. Yu, B. Yuan and Y. Feng, Talanta, 2012, 90, 123–131.
11. J. J. Rios, A. Morales and G. Marquez-Ruiz, Talanta, 2010, 80, 2076–2082.
12. Y. Guo and K. Kannan, Environ. Sci. Technol., 2013, 47, 14442–14449.
13. H. Zhao, M. Huang, J. Wu, L. Wang and H. He, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2016, 1011, 33–44.
14. A. Mousa, C. Basheer and A. Rahman Al-Arfaj, J. Sep. Sci., 2013, 36, 2003–2009.
15. M. Bonoli, P. Colabufalo, M. Pelillo, T. Gallina and G. Lercker, J. Agric. Food Chem., 2003, 51, 1141–1147.
16. N. Hayeck, S. Ravier, R. Gemayel, S. Gligorovski, I. Poulet, J. Maalouly and H. Wortham, Talanta, 2015, 144, 1163–1170.
17. M. I. Leong, M. R. Fuh and S. D. Huang, J. Chromatogr. A, 2014, 1335, 2–14.
18. M. Rezaee, Y. Assadi, M. R. Milani Hosseini, E. Aghaee, F. Ahmadi and S. Berijani, J. Chromatogr. A, 2006, 1116, 1–9.
19. G. D'Orazio, M. Asensio-Ramos, J. Hernandez-Borges, S. Fanali and M. A. Rodriguez-Delgado, J. Chromatogr. A, 2014, 1344, 109–121.
20. S. Pedersen-Bjergaard and K. E. Rasmussen, Anal. Chem., 1999, 71, 2650–2656.
21. P. Li, B. Hu, M. He and B. Chen, J. Chromatogr. A, 2014, 1356, 23–31.
22. H. Liu and P. K. Dasgupta, Anal. Chem., 1996, 68, 1817–1821.
23. H. R. Lee, S. M. Cho, J. Kim and D. S. Chung, J. Chromatogr. A, 2014, 1346, 117–122.
24. M. R. Khalili Zanjani, Y. Yamini, S. Shariati and J. A. Jonsson, Anal. Chim. Acta, 2007, 585, 286–293.
25. H. Farahani, Y. Yamini, S. Shariati, M. R. Khalili-Zanjani and S. Mansour-Baghahi, Anal. Chim. Acta, 2008, 626, 166–173.
26. V. Vickackaite and E. Pusvaskiene, J. Sep. Sci., 2009, 32, 3512–3520.
27. M. Safdarian, P. Hashemi and M. Naderlou, J. Chromatogr. A, 2012, 1244, 14–19.
28. M. I. Leong and S. D. Huang, J. Chromatogr. A, 2008, 1211, 8–12.
29. T. Asadollahi, S. Dadfarnia and A. M. Shabani, Talanta, 2010, 82, 208–212.
30. M. I. Leong and S. D. Huang, J. Chromatogr. A, 2009, 1216, 7645–7650.
31. H. Yan and H. Wang, J. Chromatogr. A, 2013, 1295, 1–15.
32. M. Asensio-Ramos, J. Hernandez-Borges, T. M. Borges-Miquel and M. A. Rodriguez-Delgado, J. Chromatogr. A, 2011, 1218, 4808–4816.
33. A. Berthod, M. J. Ruiz-Angel and S. Carda-Broch, J. Chromatogr. A, 2008, 1184, 6–18.
34. Y. Zhang and H. K. Lee, Anal. Chim. Acta, 2012, 750, 120–126.
35. L. Vidal, M. L. Riekkola and A. Canals, Anal. Chim. Acta, 2012, 715, 19–41.
36. Z. Li, Y. Pei, H. Wang, J. Fan and J. Wang, TrAC, Trends Anal. Chem., 2010, 29, 1336–1346.
37. D. Han, B. Tang, Y. R. Lee and K. H. Row, J. Sep. Sci., 2012, 35, 2949–2961.
38. J. G. Huddleston, H. D. Willauer, R. P. Swatloski, A. E. Visser and R. D. Rogers, Chem. Commun., 1998, 16, 1765–1766.
39. H. Abdolmohammad-Zadeh and G. H. Sadeghi, Talanta, 2010, 81, 778–785.
40. L. Vallecillos, E. Pocurull and F. Borrull, Talanta, 2012, 99, 824–832.
41. W. Liu, Z. Wei, Q. Zhang, F. Wu, Z. Lin, Q. Lu, F. Lin, G. Chen and L. Zhang, Talanta, 2012, 88, 43–49.
42. M. D. Joshi and J. L. Anderson, RSC Adv., 2012, 2, 5470–5484.
43. S. Anja, K. Daniel, F. Lisa, H. Stephan, G. Markus, G. Michael, K. Regina and B. K. Keppler, Aust. J. Chem., 2010, 63, 511–524.
44. K. J. Fraser and D. R. MacFarlane, Aust. J. Chem., 2009, 62, 309–321.
45. J. P. Mikkola, P. Virtanen and R. Sjöholm, Green Chem., 2006, 8, 250–255.
46. K. Yoshii, K. Yamaji, T. Tsuda, K. Tsunashima, H. Yoshida, M. Ozaki and S. Kuwabata, J. Phys. Chem. B, 2013, 117, 15051–15059.
47. K. Tsunashima and M. Sugiya, Electrochemistry, 2007, 9, 734–736.
48. H. R. Sobhi, Y. Yamini, A. Esrafili and M. Adib, J. Pharm. Biomed. Anal., 2008, 48, 1059–1063.
49. M. Li, J. Zhang, Y. Li, B. Peng, W. Zhou and H. Gao, Talanta, 2013, 107, 81–87.
50. M. Yang, P. Zhang, L. Hu, R. Lu, W. Zhou, S. Zhang and H. Gao, J. Chromatogr. A, 2014, 1360, 47–56.
51. H. Wu, H. Tian, M. F. Chen, J. C. You, L. M. Du and Y. L. Fu, J. Agric. Food Chem., 2014, 62, 7682–7689.
52. P. Liang, J. Xu and Q. Li, Anal. Chim. Acta, 2008, 609, 53–58.
53. A. Mohammadi, R. Tavakoli, M. Kamankesh, H. Rashedi, A. Attaran and M. Delavar, Anal. Chim. Acta, 2013, 804, 104–110.
54. L. B. Escudero, A. Castro Grijalba, E. M. Martinis and R. G. Wuilloud, Anal. Bioanal. Chem., 2013, 405, 7597–7613.
55. L. Yangcheng, L. Quan, L. Guangsheng and D. Youyuan, Anal. Chim. Acta, 2006, 566, 259–264.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00788k

---