2757
J. Sep. Sci. 2014, 37, 2757–2763
Tao Lv1,2 Xian-En Zhao1,2 Shuyun Zhu1,2 Fei Qu1,2 Cuihua Song1,2 Jinmao You1,2,3 ∗ Yourui Suo3 1 Shandong
Provincial Key Laboratory of Life-Organic Analysis, College of Chemistry and Chemical Engineering, Qufu Normal University, Qufu, P.R. China 2 Key Laboratory of Pharmaceutical Intermediates and Analysis of Natural Medicine, P.R. China 3 Key Laboratory of Tibetan Medicine Research, Northwest Institute of Plateau Biology, Chinese Academy of Science, Qinghai, P.R. China Received June 11, 2014 Revised July 6, 2014 Accepted July 10, 2014
Research Article
Determination of bisphenol A, 4-octylphenol, and 4-nonylphenol in soft drinks and dairy products by ultrasound-assisted dispersive liquid–liquid microextraction combined with derivatization and high-performance liquid chromatography with fluorescence detection A novel hyphenated method based on ultrasound-assisted dispersive liquid–liquid microextraction coupled to precolumn derivatization has been established for the simultaneous determination of bisphenol A, 4-octylphenol, and 4-nonylphenol by high-performance liquid chromatography with fluorescence detection. Different parameters that influence microextraction and derivatization have been optimized. The quantitative linear range of analytes is 5.0–400.0 ng/L, and the correlation coefficients are more than 0.9998. Limits of detection for soft drinks and dairy products have been obtained in the range of 0.5–1.2 ng/kg and 0.01–0.04 g/kg, respectively. Relative standard deviations of intra- and inter-day precision for retention time and peak area are in the range of 0.47–2.31 and 2.76–8.79%, respectively. Accuracy is satisfactory in the range of 81.5–118.7%. Relative standard deviations of repeatability are in the range of 0.35–1.43 and 2.36–4.75% for retention time and peak area, respectively. Enrichment factors for bisphenol A, 4-octylphenol, and 4-nonylphenol are 170.5, 240.3, and 283.2, respectively. The results of recovery and matrix effect are in the range of 82.7–114.9 and 92.0–109.0%, respectively. The proposed method has been applied to the determination of bisphenol A, 4-octylphenol, and 4-nonylphenol in soft drinks and dairy products with much higher sensitivity than many other methods. Keywords: Derivatization / Endocrine disruptors / Food analysis / Fluorescence detection / Microextraction DOI 10.1002/jssc.201400612
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Endocrine disrupting compounds (EDCs) can interfere with the function of the endocrine system in living organisms and result in a wide range of diseases [1]. Exposure to very low concentrations of these compounds can be detrimental to health. Bisphenol A (BPA) and alkylphenols (APs) including 4-octylphenol (OP) and 4-nonylphenol (NP) are such endocrine disruptors that have been widely used in industrial Correspondence: Dr. Xian-En Zhao, Shandong Provincial Key Laboratory of Life-Organic Analysis, College of Chemistry and Chemical Engineering, Qufu Normal University, Qufu 273165, P.R. China E-mail: [email protected] Fax: +86-537-4456305
Abbreviations: AP, alkylphenol; BPA, bisphenol A; EASC, 10ethylacridone-2-sulfonyl chloride; EDC, endocrine disrupting compound; EF, enrichment factor; EFS, European Food Safety Authority; FLD, fluorescence detection; NP, 4-nonylphenol; OP, 4-octylphenol; TDI, tolerable daily intake; UA-DLLME, ultrasound-assisted dispersive liquid–liquid microextraction C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
manufacture [2–4]. Humans are easily exposed to BPA, OP, and NP via foods and drinks. Many researchers have reported that BPA, OP, and NP were detected in various kinds of food or drink products in different countries [5–9]. Because of their potential activity of endocrine disruption, it is necessary to establish a sensitive, rapid, and validated method for their determination in foods and drinks. Hitherto, BPA, OP, and NP have been analyzed by a wide range of chromatographic analysis techniques. GC–MS and LC–MS have been applied more frequently among these techniques [2,10–13]. Other methods such as LC coupled with fluorescence or electrochemical detection, as well as immunochemical methods, have also been employed [14–17]. Since these analytes show very low fluorescence and UV absorption, direct LC with fluorescence and UV detection could not meet the requirements of quantitative analysis. In addition, ∗ Additional corresponding author: Professor Jinmao You, E-mail: [email protected] Colour Online: See the article online to view Fig. 2 in colour.
www.jss-journal.com
2758
J. Sep. Sci. 2014, 37, 2757–2763
T. Lv et al.
their ionization efficiency is generally low, which restricts the sensitivity of MS detection methods. A derivatization strategy can be applied to improve these problems because of its excellent sensitivity, selectivity, and validity. In this study, a precolumn derivatization method using 10-ethylacridone2-sulfonyl chloride (EASC) as the labeling reagent followed by HPLC with fluorescence detection (FLD) has been established. Since the concentrations of aliphatic amines and EDCs in wastewater are relatively large, the EASC derivatization and HPLC–FLD procedure could sufficiently meet the demand of the quantification without particular extraction or enrichment procedures [18,19]. However, the sensitivity and accuracy are insufficient when derivatization is used alone in food samples because of the lower concentrations of EDCs and more serious matrix interference. Therefore, it is necessary to develop a combined strategy of efficient extraction and derivatization, which can meet the requirement for the determination of BPA, OP, and NP in food samples. Soft drinks and dairy products are very complex matrices and the determination of BPA, OP, and NP in low concentrations within them is a challenge. Therefore, the analyte enrichment and sample cleanup process are essential. To date, many sample pretreatment techniques have been developed for this purpose such as LLE [20], SPE [21], SPME [22], LPME [23], single-drop microextraction (SDME) [24], and liquid–liquid–liquid microextraction (LLLME) [25]. The most commonly reported sample pretreatment techniques for BPA, OP, and NP are LLE and SPE. However, they demand a large amount of samples and extraction solvents, which limit their applications. Dispersive liquid–liquid microextraction (DLLME) first reported by Rezaee is a novel sample pretreatment technique [26, 27]. DLLME has many obvious advantages such as simplicity, easy operation, rapidity, cost effectiveness, slightly matrix effect, and high enrichment factor and recovery [28, 29]. Recently, ultrasoundassisted DLLME (UA-DLLME) has attracted much attention due to its speediness and efficiency [30, 31], because ultrasonic radiation is an efficient method to accelerate the mass transfer process. In this work, a novel hyphenated technique has been established for the simultaneous determination of BPA, OP, and NP in soft drinks and dairy products by the combination of UA-DLLME and derivatization. The experimental conditions of UA-DLLME and derivatization are optimized by the single-factor analysis method. To the best of our knowledge, this is the first report about the combined use of UA-DLLME and derivatization for the analysis of BPA, OP, and NP by HPLC–FLD.
2 Materials and methods 2.1 Reagents and materials OP, NP, and BPA standards were all obtained from Dr. Ehrenstorfer (Germany) with purity higher than 98%. Acetonitrile, formic acid, and other reagents were purchased from Shanghai Chemical Reagent (China). Acetonitrile and formic acid C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
were of HPLC grade, other reagents were of analytical grade. Milli-Q water purification system (USA) was applied to produce ultrapure water. EASC was synthesized as previously described in our laboratory [14]. Bottled cola and green tea as well as milk and milk powder were purchased from a local store in Qufu City. Individual stock solutions of three kinds of analytes were prepared in acetonitrile at 2 × 10−2 M. The mixed standard solutions were prepared by mixing each of the individual stock solutions and diluted with acetonitrile. The derivatization reagent solution (1.0 × 10−3 M) was prepared by dissolving 3.21 mg EASC in 10 mL acetonitrile. When not in use, all the solutions were stored at 4⬚C in a refrigerator.
2.2 Instrumentation HPLC separation and analysis were performed using Agilent 1100 HPLC systems (USA). An automatic electronic water bath (China), a Xiangzhi TGL16M high-speed refrigerated centrifuge (China), a KQ2200E ultrasonic cleaner (China), and a VX-200 vortex mixer (Labnet, USA) were equipped for derivatization and UA-DLLME experiments.
2.3 UA-DLLME 2.3.1 UA-DLLME for cola and green tea Before analysis, cola samples were degassed by ultrasonication. After that, 3 mL of soft drink sample was added into a 5 mL centrifuge tube. Then a mixture of 80 L of dichloromethane and 250 L of acetone was injected into the soft drink sample rapidly to form a cloudy solution. After treatment by ultrasound for 120 s, the mixture solution was centrifuged for 1.5 min at 12 000 rpm. Then the organic phase was gathered at the bottom of the centrifuge tube and aspirated into a syringe. The collected organic phase was transferred to a vial and dried under nitrogen for the next derivatization procedure.
2.3.2 UA-DLLME for milk and milk powder Deproteinization was the first step for HPLC analysis of dairy products. A volume of 6 mL milk sample and 1000 L perchloric acid solution (30%) were added in a 10 mL centrifuge tube. After vortexing for 10 s, the mixture solution was placed into an ultrasonic bath and shaken for 5 min followed by centrifugation at 10 000 rpm for 5 min. Three milliliters of supernate was transferred to a 5 mL centrifuge tube for UA-DLLME. In the same way, 1 g milk powder sample was dissolved in 6 mL ultrapure water in a 10 mL centrifuge tube. The deproteinization process was the same as above of the milk sample. Other operations of UA-DLLME for milk and milk powder were similar to the above. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2757–2763
Sample Preparation
2.4 Derivatization
3 Results and discussion
The reaction scheme of derivatization is shown in Supporting Information Fig. S1. A certain amount of standard or sample redissolved in acetonitrile was transferred to a 2 mL vial, then 300 L of EASC solution and 100 L of sodium bicarbonate buffer (pH 10.8) were added in order, and the mixture was shaken for about 20 s to ensure good mixing. The vial was sealed and heated for 3 min in a water bath at 50⬚C. After cooling, 20 L 25% acetic acid solution was added to neutralize the pH close to 7.0. Derivatives were filtered through 0.22 m membrane and injected for HPLC analysis.
3.1 Optimization of UA-DLLME
2.5 HPLC conditions Derivatives were separated on an RP Agilent Zorbax SB-C18 column (4.6 × 150 mm, 5 m, Agilent) by a gradient elution. Eluent A was 30% acetonitrile v/v and B was acetonitrile. Both of them contained 0.1% of formic acid. The gradient elution program was as follows: 0 min = 30% B; 3–10 min = 100% B. Injection volume was 10 L. The flow rate was constant at 1.0 mL/min and the column temperature was set at 30⬚C. The excitation and emission wavelengths of fluorescence detector were set at ex = 270 nm and em = 430 nm, respectively.
2759
3.1.1 Type and volume of extraction solvent The selection of extraction solvent was the key problem for the UA-DLLME procedure. Generally, the density of extraction solvent was higher than that of aqueous phase. Thus, the droplets of extraction solvent could accumulate in the bottom of the conical tube after centrifugation. In this study, five kinds of high-density solvents were optimized as extraction solvents, namely trichloroethylene, tetrachloroethylene, chlorobenzene, dichloromethane, and chloroform. A series of experiments was designed to find out the best extraction solvent. Three milliliters of ultrapure water sample spiked with 200.0 ng/L of analytes was transferred to a centrifuge tube, a mixture of 100 L extraction solvent and 500 L acetonitrile (disperser solvent) was injected into the sample. As shown in Fig. 1A, dichloromethane showed the best extraction efficiency. Further, the volume of dichloromethane was also optimized by changing its consumption in 50–200 L, when the volume of acetonitrile was constant at 500 L. Figure 1B indicated that the peak areas of analytes tended to decrease with the increase of the extraction solvent volume. Therefore, 80 L of dichloromethane was eventually selected as the optimized type and volume of extraction solvent.
3.1.2 Type and volume of disperser solvent 2.6 Method validation The linearity, sensitivity, precision, accuracy, repeatability, recovery, and matrix effect were evaluated. A series of calibration samples were prepared in the range of 5.0–400.0 ng/L. Calibration curves of analytes were constructed by plotting the peak areas of BPA, OP, and NP versus their concentrations. LOD and LOQ were used for describing the sensitivity of the method, which were calculated when the S/N was above 3:1 and 10:1, respectively. Precision and accuracy were evaluated at three concentration levels (10.0, 50.0, and 300.0 ng/L). Six replicates of each concentration level were analyzed three times over three days in order to determine the intra- and inter-day precision and accuracy using freshly prepared calibration curves. The recovery was evaluated by spiking specified amount of standards at three concentration levels (10.0, 50.0, and 300.0 ng/L) to blank green tea and milk samples. The method repeatability was investigated by measuring the RSDs of peak areas and retention times for BPA, OP, and NP under identical conditions. The matrix effect was assessed by comparing the peak areas of spiked BPA, OP, and NP in soft drink and dairy product samples using UA-DLLME and derivatization procedure to those of equivalent concentration in standard solution. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The selection of disperser solvent mainly depended on its miscibility with extraction solvent and aqueous phase. Disperser solvent played an important role in the formation of a cloudy solution and achieved a rapid equilibrium for extraction. In this article, four kinds of disperser solvent were used for the optimization of UA-DLLME, namely, acetone, acetonitrile, methanol, and ethanol. A series of experiments was designed to find out the best disperser solvent among them. The extraction solvent (dichloromethane 80 L) mixed with 500 L of varied disperser solvents as above. Other operations were similar to those in previous reports. As can be seen from Fig. 1C, when acetone was used as disperser solvent the peak areas of BPA, OP, and NP were the highest. Thus, acetone was selected as disperser solvent in this study. Similarly, the volume of acetone was also optimized. Other optimized parameters were remained the same, the volume of acetone was changed from 80 to 600 L. The peak areas increased in the range of 80–250 L of disperser solvent and then decreased along with the gradually increasing volume of acetone. A volume of 250 L acetone showed the best extraction efficiency Fig. 1D. 3.1.3 Effect of pH and ultrasonication time The pH value of the aqueous sample was also investigated. A series of experiments was carried out by altering the pH www.jss-journal.com
2760
T. Lv et al.
J. Sep. Sci. 2014, 37, 2757–2763
Figure 1. Optimization of different parameters of DLLME at 200.0 ng/L of BPA, OP, and NP; (A) type of extraction solvent; (B) volume of extraction solvent; (C) type of disperser solvent; and (D) volume of disperser solvent.
from 2 to 11. The peak areas of three analytes were higher in acidic aqueous solution than in alkaline aqueous solution. Moreover, a little change of peak areas from three analytes was observed in the pH range of 2–7. Therefore, all aqueous real samples were adjusted between pH 5–7 in this study. Ultrasonication could observably promote mass transfer of analytes from aqueous phase into the extraction solvent. The effect of ultrasonication time was evaluated in the range of 0–150 s and each interval was 30 s. Other optimized parameters remained the same. Along with the increasing of ultrasonication time from 0 to 120 s, peak areas of BPA, OP, and NP increased. However, peak areas of their derivatives decreased when ultrasonication time exceeded 120 s. This could be caused by the alternate high-pressure (compression) and low-pressure (rarefaction) cycles produced by ultrasonication. With the ultrasonication time increasing, many small vacuum bubbles or void were formed. When these small vacuum bubbles or void violently collapsed, the local temperature rose rapidly [32]. This influenced the extraction efficiency of UA-DLLME. Therefore, 120 s was selected as the optimum ultrasonication time for the UA-DLLME procedure.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.2 Optimization of derivatization conditions The main factors such as amount of EASC, the pH value of the sodium bicarbonate buffer, reaction time, and temperature were all optimized. BPA had two phenolic hydroxyl groups that could be labeled by EASC. But, the amount of EASC should be sufficient enough to ensure the complete derivatization of two phenolic hydroxyl groups in BPA. The amount of EASC, described as the molar ratios of EASC to BPA, OP, and NP, was optimized in the range of 3–20. The results indicated that the optimal molar ratio of EASC to analytes was 15 (Supporting Information Fig. S2A). The influence of pH on the derivatization was investigated with sodium bicarbonate buffer (0.1 M) in the pH range of 9–11.5. As can be seen from Supporting Information Fig. S2B, the maximum derivatization yield was obtained at pH 10.8. The effect of reaction temperature and time indicated that complete reaction of analytes with EASC needed 3 min at 50⬚C. When the derivatization procedure was carried out under the above optimum conditions, the hydroxyl group of BPA, OP, and NP could be sufficiently labeled.
www.jss-journal.com
Sample Preparation
J. Sep. Sci. 2014, 37, 2757–2763
2761
area was obtained for BPA, OP, and NP (R2 > 0.9998). For soft drink and dairy product samples analysis, the LODs of BPA, OP, and NP were in the range of 0.5–1.2 ng/L and 0.01– 0.04 g/kg, and the LOQs were in the range of 1.6–3.9 ng/L and 0.03–0.12 g/kg, respectively (Table 1). 3.4.2 Accuracy, precision, and repeatability As can be seen from Supporting Information Table S1, accuracies were in the range of 81.5–118.7%. Precision expressed as RSDs of retention time and peak area was in the range of 0.47–2.31% and 2.76–8.79%, respectively. RSDs values of repeatability were in the range of 0.35–1.43% and 2.36–4.75% for retention time and peak area, respectively. In conclusion, good accuracy, precision, and repeatability were obtained for the proposed method. 3.4.3 Recovery and matrix effect The mean recoveries of BPA, OP, and NP for green tea and milk samples were in the range of 84.3–114.9% and 82.7– 112.5%, respectively (Supporting Information Table S1). The mean values of matrix effect for four kinds of real samples were in the range of 92.0–109.0% (Supporting Information Table S2), which indicated that the novel hyphenated technique of UA-DLLME and derivatization could effectively improve the matrix effect. This might be because UA-DLLME is a selective extraction procedure for BPA, OP, and NP in complex matrix samples, and EASC derivatization was also a selective labeling procedure for the hydroxyl group in three analytes. Figure 2. Chromatograms of BPA, OP, and NP derivatives from (A) standard derivatives of BPA, OP, and NP; (B) a green tea sample; and (C) a milk powder sample with their spiked ones.
3.3 HPLC separation Various analytical columns, mobile-phase compositions and flow rates were investigated to obtain satisfactory separation. Complete HPLC separation of the derivatives was achieved on an Agilent Zorbax SB-C18 column with gradient elution using eluent A acetonitrile/H2 O (30:70, v/v) and eluent B acetonitrile. The peak shape and resolution of derivatives were improved when 0.1% formic acid was added to the mobile phase. The chromatograms of BPA, OP, and NP standard (A), green tea (B), and milk powder (C) sample and the spiked samples are shown in Fig. 2.
3.4 Method validation 3.4.1 Linearity A seven-point calibration curve for each kind of analyte was plotted in the range of 5.0–400.0 ng/L. As can be seen from Table 1, a good correlation between concentration and peak C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.5 Enrichment factors The enrichment factors (EFs) were defined as the ratio of the analytes concentration in the extracted phase (Cexp ) to the initial concentration of analytes (C0 ) in the sample: EF = Cexp /C0 . Under the optimum extraction and derivatization conditions above, results of EFs were 170.5, 240.3, and 283.2 for BPA, OP, and NP, respectively.
3.6 Comparison with reported methods In (Table 2), LODs of the proposed method were compared with those of the reported methods for the determination of BPA, OP, and NP in different kinds of real samples [2, 9, 33–38]. Compared with pressurized liquid extraction [2], SPME [34], SDME [35], and SPE [9, 36], DLLME generally showed relatively lower LODs [33, 35]. Moreover, the proposed method in this study using EASC as derivatization reagents was also compared with those of other derivatization reagents or without a derivatization procedure. The LODs of these methods with a derivatization procedure by HPLC–FLD [36–38] were relatively lower than LC–MS/MS [2], GC–MS [9], and HPLC–UV [33–35] without derivatization. www.jss-journal.com
2762
J. Sep. Sci. 2014, 37, 2757–2763
T. Lv et al.
Table 1. Linear regression equations, correlation coefficients, LODs, and LOQs of three phenol derivatives
Phenol derivatives
BPA OP NP
Linear regression equations Y = AX + Ba)
Y = 2.78X – 1.78 Y = 1.07X + 0.44 Y = 1.03X + 3.06
R2
0.9999 0.9998 0.9998
Soft drinks
Dairy products
LOD (ng/L)
LOQ (ng/L)
LOD (g/kg)
LOQ (g/kg)
0.5 1.0 1.2
1.6 3.3 3.9
0.01 0.03 0.04
0.03 0.10 0.12
Detection method
LOD (g/kg)
a) X is concentration of BPA, OP, and NP (ng/L); Y is peak area. Table 2. Comparison of the proposed method with reported methods
Sample type
Extraction method
Derivatization reagent
Ref.
BPA
OP
NP
Milk powder Infant formula River and tap water Wastewater
Pressurized liquid extraction SPE DLLME SPME
– – – –
LC–MS/MS GC–MS HPLC–UV HPLC–UV
5.0 0.5 0.07 0.3
3.0 – – 0.8
5.0 – –
[2] [9] [33] [34]
Seawaters
SDME DLLME
– –
HPLC–UV HPLC–UV
4.0 0.7
9.0 0.8
9.0 1.6
[35]
Urine Sediment Plastic Milk and milk powder
SPE SPE SPE UA-DLLME
p-Nitrobenzoyl chloride CEOCa) DIB-Clb) EASC
HPLC–FLD HPLC–FLD HPLC–FLD HPLC–FLD
2.7 0.1 0.67 0.01
– 0.2 – 0.03
2.9 0.2 1.3 0.04
[36] [37] [38] This work
–, not included in the method. a) CEOC, 2-(9H-carbazol-9-yl)ethyl carbonochloridate. b) DIB-Cl, 4-(4,5-diphenyl-1H-imidazol-2-yl)benzoyl chloride.
However, LODs of this study for BPA, OP, and NP were about 7–900 times lower than the reported methods listed in (Table 2) [2, 9, 33–38]. This should be the benefit of the use of the hyphenated technique of UA-DLLME and derivatization followed by HPLC and fluorescence detection, which was not reported for the determination of BPA, OP, and NP. In conclusion, the proposed strategy of UA-DLLME combined with EASC derivatization could greatly increase the sensitivity and the selectivity, and the matrix effect was also effectively improved.
3.7 Application to real samples This method was successfully applied for the determination of BPA, OP, and NP in different soft drinks and dairy products. In this study, three kinds of different brands of cola, green tea, milk, and milk powder were used as real samples for analysis. Figure 2B and C shows the representative chromatograms of a green tea and a milk powder sample with their spiked ones. All the samples were analyzed in triplicate and the mean values were reported as the results (Supporting Information Table S2). For six kinds of soft drink sam C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ples, BPA was found with concentrations ranging from 5.3 to 78.9 ng/L; OP was found in a cola sample and a green tea sample with concentrations at 20.4 and 7.2 ng/L; NP was found in two cola samples with concentrations at 18.5 and 5.8 ng/L, respectively. For six kinds of dairy product samples, BPA was detectable in two milk powder samples and one milk sample, at levels of 0.8, 8.4, and 128.7 g/kg; OP and NP were found in two milk samples and one milk powder sample of the same at levels of 1.5, 0.3, and 29.5 g/kg and 13.7, 3.5, and 36.7 g/kg, respectively. According to the calculation method for the tolerable daily intake (TDI) of the European Food Safety Authority (EFSA) [2], the results in this study showed that BPA, OP, and NP were present in soft drinks and dairy products but levels were very low within the limits of the TDI.
4 Concluding remarks In this study, a highly sensitive and selective UA-DLLME pretreatment technique coupled with derivatization followed by HPLC–FLD was developed and validated for the determination of BPA, OP, and NP in soft drinks and dairy products. The combination of UA-DLLME and derivatization www.jss-journal.com
J. Sep. Sci. 2014, 37, 2757–2763
effectively improved the sensitivity, selectivity, and matrix effect of EDCs. This hyphenated technique is the first of its kind for the analysis of BPA, OP, and NP and it could be further extended for the analysis of other homologous analytes. The authors acknowledge the financial support from the National Natural Science Foundation of China (81303179, 21275089) and the Foundation of Qufu Normal University (2012019). The authors have declared no conflict of interest.
5 References
Sample Preparation
2763
[17] Schmitz-Afonso, I., Loyo-Rosales, J. E., de la Paz Aviles, M., Rattner, B. A., Rice, C. P., J. Chromatogr. A 2003, 1010, 25–35. [18] You, J., Zhao, H., Sun, Z., Xia, L., Yan, T., Suo, Y., Li, Y., J. Sep. Sci. 2009, 32, 1351–1362. [19] Zhang, S., You, J., Sun, Z., Song, C., Ning, S., Zhao, C., Suo, Y., Microchem. J. 2012, 103, 90–96. ´ ´ [20] Zafra, A., del Olmo, M., Suarez, B., Hontoria, E., Navalon, A., Vı́lchez, J., Water Res. 2003, 37, 735–742. [21] Vega-Morales, T., Sosa-Ferrera, Z., Santana-Rodr´ıguez, J., J. Hazard. Mater. 2010, 183, 701–711. [22] Basheer, C., Parthiban, A., Jayaraman, A., Lee, H. K., Valiyaveettil, S., J. Chromatogr. A 2005, 1087, 274–282.
[1] Diamanti-Kandarakis, E., Bourguignon, J. P., Giudice, L. C., Hauser, R., Prins, G. S., Soto, A. M., Zoeller, R. T., Gore, A. C., Endocr. Rev. 2009, 30, 293–342.
[23] Kawaguchi, M., Ito, R., Endo, N., Okanouchi, N., Sakui, N., Saito, K., Nakazawa, H., J. Chromatogr. A 2006, 1110, 1–5.
˜ [2] Ferrer, E., Santoni, E., Vittori, S., Font, G., Manes, J., Sagratini, G., Food Chem. 2011, 126, 360–367.
[24] Xu, L., Basheer, C., Lee, H. K., J. Chromatogr. A 2007, 1152, 184–192.
[3] Adamusova, H., Bosakova, Z., Coufal, P., Pacakova, V., J. Sep. Sci. 2014, 37, 885–905.
[25] Lin, C. Y., Fuh, M. R., Huang, S. D., J. Sep. Sci. 2011, 34, 428–435.
[4] Mezcua, M., Martinez-Uroz, M. A., Gomez-Ramos, M. M., Gomez, M. J., Navas, J. M., Fernandez-Alba, A. R., Talanta 2012, 100, 90–106.
[26] Rezaee, M., Assadi, Y., Milani Hosseini, M.-R., Aghaee, E., Ahmadi, F., Berijani, S., J. Chromatogr. A 2006, 1116, 1–9.
[5] Fernandes, A., Costley, C., Rose, M., Food Addit. Contam. 2003, 20, 846–852.
´ ´ ´ [27] Lopez-Darias, J., German-Hern andez, M., Pino, V., Afonso, A. M., Talanta 2010, 80, 1611–1618.
[6] Guenther, K., Heinke, V., Thiele, B., Kleist, E., Prast, H., Raecker, T., Environ. Sci. Technol. 2002, 36, 1676–1680.
[28] Vinas, P., Campillo, N., Lopez-Garcia, I., HernandezCordoba, M., Anal. Bioanal. Chem. 2014, 406, 2067–2099.
[7] Schecter, A., Malik, N., Haffner, D., Smith, S., Harris, T. R., Paepke, O., Birnbaum, L., Environ. Sci. Technol. 2010, 44, 9425–9430. [8] Yonekubo, J., Hayakawa, K., Sajiki, J., J. Agric. Food Chem. 2008, 56, 2041–2047. [9] Cao, X. L., Dufresne, G., Belisle, S., Clement, G., Falicki, M., Beraldin, F., Rulibikiye, A., J. Agric. Food. Chem. 2008, 56, 7919–7924. [10] Kawaguchi, M., Inoue, K., Yoshimura, M., Ito, R., Sakui, N., Okanouchi, N., Nakazawa, H., J. Chromatogr. B 2004, 805, 41–48. [11] Jiao, Y., Ding, L., Fu, S., Zhu, S., Li, H., Wang, L., Anal Methods 2012, 4, 291–298. ´ ˜ [12] Llorca-Porcel, J., Mart´ınez-Parreno, M., Mart´ınezSoriano, E., Valor, I., J. Chromatogr. A 2009, 1216, 5955– 5961.
[29] Saraji, M., Boroujeni, M. K., Anal. Bioanal. Chem. 2014, 406, 2027–2066. [30] Parrilla Vazquez, M. M., Parrilla Vazquez, P., Martinez Galera, M., Gil Garcia, M. D., Ucles, A., J. Chromatogr. A 2013, 1291, 19–26. [31] Campone, L., Piccinelli, A. L., Pagano, I., Carabetta, S., Di Sanzo, R., Russo, M., Rastrelli, L., J. Chromatogr. A 2014, 1334, 9–15. [32] Iparraguirre, A., Prieto, A., Navarro, P., Olivares, M., ´ Fernandez, L., Zuloaga, O., Anal. Bioanal. Chem. 2011, 401, 339–352. [33] Rezaee, M., Yamini, Y., Shariati, S., Esrafili, A., Shamsipur, M., J. Chromatogr. A 2009, 1216, 1511–1514. [34] Penalver, A., Pocurull, E., Borrull, F., Marce, R., J. Chromatogr. A 2002, 964, 153–160.
ˇ ıcek, ˇ ´ A., Vlcek, ˇ [13] Matej´ D., Grycova, J., J. Sep. Sci. 2013, 36, 1097–1103.
[35] Lopez-Darias, J., German-Hernandez, M., Pino, V., Afonso, A. M., Talanta 2010, 80, 1611–1618.
´ ´ [14] Ballesteros-Gomez, A., Rubio, S., Perez-Bendito, D., J. Chromatogr. A 2009, 1216, 449–469.
[36] Mao, L., Sun, C., Zhang, H., Li, Y., Wu, D., Anal. Chim. Acta 2004, 522, 241–246.
[15] Braunrath, R., Podlipna, D., Padlesak, S., Cichna-Markl, M., J. Agric. Food Chem. 2005, 53, 8911–8917.
[37] Zhang, S. J., You, J. M., Song, C. H., Chen, G., Suo, Y. R., Anal Methods 2012, 4, 4030–4036.
´ [16] Garc´ıa-Prieto, A., Lunar, M. L., Rubio, S., Perez-Bendito, D., Anal. Chim. Acta 2008, 630, 19–27.
[38] Sun, Y., Wada, M., Kuroda, N., Hirayama, K., Nakazawa, H., Nakashima, K., Anal. Sci. 2001, 17, 697–702.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2757–2763
Tao Lv1,2 Xian-En Zhao1,2 Shuyun Zhu1,2 Fei Qu1,2 Cuihua Song1,2 Jinmao You1,2,3 ∗ Yourui Suo3 1 Shandong
Provincial Key Laboratory of Life-Organic Analysis, College of Chemistry and Chemical Engineering, Qufu Normal University, Qufu, P.R. China 2 Key Laboratory of Pharmaceutical Intermediates and Analysis of Natural Medicine, P.R. China 3 Key Laboratory of Tibetan Medicine Research, Northwest Institute of Plateau Biology, Chinese Academy of Science, Qinghai, P.R. China Received June 11, 2014 Revised July 6, 2014 Accepted July 10, 2014
Research Article
Determination of bisphenol A, 4-octylphenol, and 4-nonylphenol in soft drinks and dairy products by ultrasound-assisted dispersive liquid–liquid microextraction combined with derivatization and high-performance liquid chromatography with fluorescence detection A novel hyphenated method based on ultrasound-assisted dispersive liquid–liquid microextraction coupled to precolumn derivatization has been established for the simultaneous determination of bisphenol A, 4-octylphenol, and 4-nonylphenol by high-performance liquid chromatography with fluorescence detection. Different parameters that influence microextraction and derivatization have been optimized. The quantitative linear range of analytes is 5.0–400.0 ng/L, and the correlation coefficients are more than 0.9998. Limits of detection for soft drinks and dairy products have been obtained in the range of 0.5–1.2 ng/kg and 0.01–0.04 g/kg, respectively. Relative standard deviations of intra- and inter-day precision for retention time and peak area are in the range of 0.47–2.31 and 2.76–8.79%, respectively. Accuracy is satisfactory in the range of 81.5–118.7%. Relative standard deviations of repeatability are in the range of 0.35–1.43 and 2.36–4.75% for retention time and peak area, respectively. Enrichment factors for bisphenol A, 4-octylphenol, and 4-nonylphenol are 170.5, 240.3, and 283.2, respectively. The results of recovery and matrix effect are in the range of 82.7–114.9 and 92.0–109.0%, respectively. The proposed method has been applied to the determination of bisphenol A, 4-octylphenol, and 4-nonylphenol in soft drinks and dairy products with much higher sensitivity than many other methods. Keywords: Derivatization / Endocrine disruptors / Food analysis / Fluorescence detection / Microextraction DOI 10.1002/jssc.201400612
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Endocrine disrupting compounds (EDCs) can interfere with the function of the endocrine system in living organisms and result in a wide range of diseases [1]. Exposure to very low concentrations of these compounds can be detrimental to health. Bisphenol A (BPA) and alkylphenols (APs) including 4-octylphenol (OP) and 4-nonylphenol (NP) are such endocrine disruptors that have been widely used in industrial Correspondence: Dr. Xian-En Zhao, Shandong Provincial Key Laboratory of Life-Organic Analysis, College of Chemistry and Chemical Engineering, Qufu Normal University, Qufu 273165, P.R. China E-mail: [email protected] Fax: +86-537-4456305
Abbreviations: AP, alkylphenol; BPA, bisphenol A; EASC, 10ethylacridone-2-sulfonyl chloride; EDC, endocrine disrupting compound; EF, enrichment factor; EFS, European Food Safety Authority; FLD, fluorescence detection; NP, 4-nonylphenol; OP, 4-octylphenol; TDI, tolerable daily intake; UA-DLLME, ultrasound-assisted dispersive liquid–liquid microextraction C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
manufacture [2–4]. Humans are easily exposed to BPA, OP, and NP via foods and drinks. Many researchers have reported that BPA, OP, and NP were detected in various kinds of food or drink products in different countries [5–9]. Because of their potential activity of endocrine disruption, it is necessary to establish a sensitive, rapid, and validated method for their determination in foods and drinks. Hitherto, BPA, OP, and NP have been analyzed by a wide range of chromatographic analysis techniques. GC–MS and LC–MS have been applied more frequently among these techniques [2,10–13]. Other methods such as LC coupled with fluorescence or electrochemical detection, as well as immunochemical methods, have also been employed [14–17]. Since these analytes show very low fluorescence and UV absorption, direct LC with fluorescence and UV detection could not meet the requirements of quantitative analysis. In addition, ∗ Additional corresponding author: Professor Jinmao You, E-mail: [email protected] Colour Online: See the article online to view Fig. 2 in colour.
www.jss-journal.com
2758
J. Sep. Sci. 2014, 37, 2757–2763
T. Lv et al.
their ionization efficiency is generally low, which restricts the sensitivity of MS detection methods. A derivatization strategy can be applied to improve these problems because of its excellent sensitivity, selectivity, and validity. In this study, a precolumn derivatization method using 10-ethylacridone2-sulfonyl chloride (EASC) as the labeling reagent followed by HPLC with fluorescence detection (FLD) has been established. Since the concentrations of aliphatic amines and EDCs in wastewater are relatively large, the EASC derivatization and HPLC–FLD procedure could sufficiently meet the demand of the quantification without particular extraction or enrichment procedures [18,19]. However, the sensitivity and accuracy are insufficient when derivatization is used alone in food samples because of the lower concentrations of EDCs and more serious matrix interference. Therefore, it is necessary to develop a combined strategy of efficient extraction and derivatization, which can meet the requirement for the determination of BPA, OP, and NP in food samples. Soft drinks and dairy products are very complex matrices and the determination of BPA, OP, and NP in low concentrations within them is a challenge. Therefore, the analyte enrichment and sample cleanup process are essential. To date, many sample pretreatment techniques have been developed for this purpose such as LLE [20], SPE [21], SPME [22], LPME [23], single-drop microextraction (SDME) [24], and liquid–liquid–liquid microextraction (LLLME) [25]. The most commonly reported sample pretreatment techniques for BPA, OP, and NP are LLE and SPE. However, they demand a large amount of samples and extraction solvents, which limit their applications. Dispersive liquid–liquid microextraction (DLLME) first reported by Rezaee is a novel sample pretreatment technique [26, 27]. DLLME has many obvious advantages such as simplicity, easy operation, rapidity, cost effectiveness, slightly matrix effect, and high enrichment factor and recovery [28, 29]. Recently, ultrasoundassisted DLLME (UA-DLLME) has attracted much attention due to its speediness and efficiency [30, 31], because ultrasonic radiation is an efficient method to accelerate the mass transfer process. In this work, a novel hyphenated technique has been established for the simultaneous determination of BPA, OP, and NP in soft drinks and dairy products by the combination of UA-DLLME and derivatization. The experimental conditions of UA-DLLME and derivatization are optimized by the single-factor analysis method. To the best of our knowledge, this is the first report about the combined use of UA-DLLME and derivatization for the analysis of BPA, OP, and NP by HPLC–FLD.
2 Materials and methods 2.1 Reagents and materials OP, NP, and BPA standards were all obtained from Dr. Ehrenstorfer (Germany) with purity higher than 98%. Acetonitrile, formic acid, and other reagents were purchased from Shanghai Chemical Reagent (China). Acetonitrile and formic acid C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
were of HPLC grade, other reagents were of analytical grade. Milli-Q water purification system (USA) was applied to produce ultrapure water. EASC was synthesized as previously described in our laboratory [14]. Bottled cola and green tea as well as milk and milk powder were purchased from a local store in Qufu City. Individual stock solutions of three kinds of analytes were prepared in acetonitrile at 2 × 10−2 M. The mixed standard solutions were prepared by mixing each of the individual stock solutions and diluted with acetonitrile. The derivatization reagent solution (1.0 × 10−3 M) was prepared by dissolving 3.21 mg EASC in 10 mL acetonitrile. When not in use, all the solutions were stored at 4⬚C in a refrigerator.
2.2 Instrumentation HPLC separation and analysis were performed using Agilent 1100 HPLC systems (USA). An automatic electronic water bath (China), a Xiangzhi TGL16M high-speed refrigerated centrifuge (China), a KQ2200E ultrasonic cleaner (China), and a VX-200 vortex mixer (Labnet, USA) were equipped for derivatization and UA-DLLME experiments.
2.3 UA-DLLME 2.3.1 UA-DLLME for cola and green tea Before analysis, cola samples were degassed by ultrasonication. After that, 3 mL of soft drink sample was added into a 5 mL centrifuge tube. Then a mixture of 80 L of dichloromethane and 250 L of acetone was injected into the soft drink sample rapidly to form a cloudy solution. After treatment by ultrasound for 120 s, the mixture solution was centrifuged for 1.5 min at 12 000 rpm. Then the organic phase was gathered at the bottom of the centrifuge tube and aspirated into a syringe. The collected organic phase was transferred to a vial and dried under nitrogen for the next derivatization procedure.
2.3.2 UA-DLLME for milk and milk powder Deproteinization was the first step for HPLC analysis of dairy products. A volume of 6 mL milk sample and 1000 L perchloric acid solution (30%) were added in a 10 mL centrifuge tube. After vortexing for 10 s, the mixture solution was placed into an ultrasonic bath and shaken for 5 min followed by centrifugation at 10 000 rpm for 5 min. Three milliliters of supernate was transferred to a 5 mL centrifuge tube for UA-DLLME. In the same way, 1 g milk powder sample was dissolved in 6 mL ultrapure water in a 10 mL centrifuge tube. The deproteinization process was the same as above of the milk sample. Other operations of UA-DLLME for milk and milk powder were similar to the above. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2757–2763
Sample Preparation
2.4 Derivatization
3 Results and discussion
The reaction scheme of derivatization is shown in Supporting Information Fig. S1. A certain amount of standard or sample redissolved in acetonitrile was transferred to a 2 mL vial, then 300 L of EASC solution and 100 L of sodium bicarbonate buffer (pH 10.8) were added in order, and the mixture was shaken for about 20 s to ensure good mixing. The vial was sealed and heated for 3 min in a water bath at 50⬚C. After cooling, 20 L 25% acetic acid solution was added to neutralize the pH close to 7.0. Derivatives were filtered through 0.22 m membrane and injected for HPLC analysis.
3.1 Optimization of UA-DLLME
2.5 HPLC conditions Derivatives were separated on an RP Agilent Zorbax SB-C18 column (4.6 × 150 mm, 5 m, Agilent) by a gradient elution. Eluent A was 30% acetonitrile v/v and B was acetonitrile. Both of them contained 0.1% of formic acid. The gradient elution program was as follows: 0 min = 30% B; 3–10 min = 100% B. Injection volume was 10 L. The flow rate was constant at 1.0 mL/min and the column temperature was set at 30⬚C. The excitation and emission wavelengths of fluorescence detector were set at ex = 270 nm and em = 430 nm, respectively.
2759
3.1.1 Type and volume of extraction solvent The selection of extraction solvent was the key problem for the UA-DLLME procedure. Generally, the density of extraction solvent was higher than that of aqueous phase. Thus, the droplets of extraction solvent could accumulate in the bottom of the conical tube after centrifugation. In this study, five kinds of high-density solvents were optimized as extraction solvents, namely trichloroethylene, tetrachloroethylene, chlorobenzene, dichloromethane, and chloroform. A series of experiments was designed to find out the best extraction solvent. Three milliliters of ultrapure water sample spiked with 200.0 ng/L of analytes was transferred to a centrifuge tube, a mixture of 100 L extraction solvent and 500 L acetonitrile (disperser solvent) was injected into the sample. As shown in Fig. 1A, dichloromethane showed the best extraction efficiency. Further, the volume of dichloromethane was also optimized by changing its consumption in 50–200 L, when the volume of acetonitrile was constant at 500 L. Figure 1B indicated that the peak areas of analytes tended to decrease with the increase of the extraction solvent volume. Therefore, 80 L of dichloromethane was eventually selected as the optimized type and volume of extraction solvent.
3.1.2 Type and volume of disperser solvent 2.6 Method validation The linearity, sensitivity, precision, accuracy, repeatability, recovery, and matrix effect were evaluated. A series of calibration samples were prepared in the range of 5.0–400.0 ng/L. Calibration curves of analytes were constructed by plotting the peak areas of BPA, OP, and NP versus their concentrations. LOD and LOQ were used for describing the sensitivity of the method, which were calculated when the S/N was above 3:1 and 10:1, respectively. Precision and accuracy were evaluated at three concentration levels (10.0, 50.0, and 300.0 ng/L). Six replicates of each concentration level were analyzed three times over three days in order to determine the intra- and inter-day precision and accuracy using freshly prepared calibration curves. The recovery was evaluated by spiking specified amount of standards at three concentration levels (10.0, 50.0, and 300.0 ng/L) to blank green tea and milk samples. The method repeatability was investigated by measuring the RSDs of peak areas and retention times for BPA, OP, and NP under identical conditions. The matrix effect was assessed by comparing the peak areas of spiked BPA, OP, and NP in soft drink and dairy product samples using UA-DLLME and derivatization procedure to those of equivalent concentration in standard solution. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The selection of disperser solvent mainly depended on its miscibility with extraction solvent and aqueous phase. Disperser solvent played an important role in the formation of a cloudy solution and achieved a rapid equilibrium for extraction. In this article, four kinds of disperser solvent were used for the optimization of UA-DLLME, namely, acetone, acetonitrile, methanol, and ethanol. A series of experiments was designed to find out the best disperser solvent among them. The extraction solvent (dichloromethane 80 L) mixed with 500 L of varied disperser solvents as above. Other operations were similar to those in previous reports. As can be seen from Fig. 1C, when acetone was used as disperser solvent the peak areas of BPA, OP, and NP were the highest. Thus, acetone was selected as disperser solvent in this study. Similarly, the volume of acetone was also optimized. Other optimized parameters were remained the same, the volume of acetone was changed from 80 to 600 L. The peak areas increased in the range of 80–250 L of disperser solvent and then decreased along with the gradually increasing volume of acetone. A volume of 250 L acetone showed the best extraction efficiency Fig. 1D. 3.1.3 Effect of pH and ultrasonication time The pH value of the aqueous sample was also investigated. A series of experiments was carried out by altering the pH www.jss-journal.com
2760
T. Lv et al.
J. Sep. Sci. 2014, 37, 2757–2763
Figure 1. Optimization of different parameters of DLLME at 200.0 ng/L of BPA, OP, and NP; (A) type of extraction solvent; (B) volume of extraction solvent; (C) type of disperser solvent; and (D) volume of disperser solvent.
from 2 to 11. The peak areas of three analytes were higher in acidic aqueous solution than in alkaline aqueous solution. Moreover, a little change of peak areas from three analytes was observed in the pH range of 2–7. Therefore, all aqueous real samples were adjusted between pH 5–7 in this study. Ultrasonication could observably promote mass transfer of analytes from aqueous phase into the extraction solvent. The effect of ultrasonication time was evaluated in the range of 0–150 s and each interval was 30 s. Other optimized parameters remained the same. Along with the increasing of ultrasonication time from 0 to 120 s, peak areas of BPA, OP, and NP increased. However, peak areas of their derivatives decreased when ultrasonication time exceeded 120 s. This could be caused by the alternate high-pressure (compression) and low-pressure (rarefaction) cycles produced by ultrasonication. With the ultrasonication time increasing, many small vacuum bubbles or void were formed. When these small vacuum bubbles or void violently collapsed, the local temperature rose rapidly [32]. This influenced the extraction efficiency of UA-DLLME. Therefore, 120 s was selected as the optimum ultrasonication time for the UA-DLLME procedure.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.2 Optimization of derivatization conditions The main factors such as amount of EASC, the pH value of the sodium bicarbonate buffer, reaction time, and temperature were all optimized. BPA had two phenolic hydroxyl groups that could be labeled by EASC. But, the amount of EASC should be sufficient enough to ensure the complete derivatization of two phenolic hydroxyl groups in BPA. The amount of EASC, described as the molar ratios of EASC to BPA, OP, and NP, was optimized in the range of 3–20. The results indicated that the optimal molar ratio of EASC to analytes was 15 (Supporting Information Fig. S2A). The influence of pH on the derivatization was investigated with sodium bicarbonate buffer (0.1 M) in the pH range of 9–11.5. As can be seen from Supporting Information Fig. S2B, the maximum derivatization yield was obtained at pH 10.8. The effect of reaction temperature and time indicated that complete reaction of analytes with EASC needed 3 min at 50⬚C. When the derivatization procedure was carried out under the above optimum conditions, the hydroxyl group of BPA, OP, and NP could be sufficiently labeled.
www.jss-journal.com
Sample Preparation
J. Sep. Sci. 2014, 37, 2757–2763
2761
area was obtained for BPA, OP, and NP (R2 > 0.9998). For soft drink and dairy product samples analysis, the LODs of BPA, OP, and NP were in the range of 0.5–1.2 ng/L and 0.01– 0.04 g/kg, and the LOQs were in the range of 1.6–3.9 ng/L and 0.03–0.12 g/kg, respectively (Table 1). 3.4.2 Accuracy, precision, and repeatability As can be seen from Supporting Information Table S1, accuracies were in the range of 81.5–118.7%. Precision expressed as RSDs of retention time and peak area was in the range of 0.47–2.31% and 2.76–8.79%, respectively. RSDs values of repeatability were in the range of 0.35–1.43% and 2.36–4.75% for retention time and peak area, respectively. In conclusion, good accuracy, precision, and repeatability were obtained for the proposed method. 3.4.3 Recovery and matrix effect The mean recoveries of BPA, OP, and NP for green tea and milk samples were in the range of 84.3–114.9% and 82.7– 112.5%, respectively (Supporting Information Table S1). The mean values of matrix effect for four kinds of real samples were in the range of 92.0–109.0% (Supporting Information Table S2), which indicated that the novel hyphenated technique of UA-DLLME and derivatization could effectively improve the matrix effect. This might be because UA-DLLME is a selective extraction procedure for BPA, OP, and NP in complex matrix samples, and EASC derivatization was also a selective labeling procedure for the hydroxyl group in three analytes. Figure 2. Chromatograms of BPA, OP, and NP derivatives from (A) standard derivatives of BPA, OP, and NP; (B) a green tea sample; and (C) a milk powder sample with their spiked ones.
3.3 HPLC separation Various analytical columns, mobile-phase compositions and flow rates were investigated to obtain satisfactory separation. Complete HPLC separation of the derivatives was achieved on an Agilent Zorbax SB-C18 column with gradient elution using eluent A acetonitrile/H2 O (30:70, v/v) and eluent B acetonitrile. The peak shape and resolution of derivatives were improved when 0.1% formic acid was added to the mobile phase. The chromatograms of BPA, OP, and NP standard (A), green tea (B), and milk powder (C) sample and the spiked samples are shown in Fig. 2.
3.4 Method validation 3.4.1 Linearity A seven-point calibration curve for each kind of analyte was plotted in the range of 5.0–400.0 ng/L. As can be seen from Table 1, a good correlation between concentration and peak C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.5 Enrichment factors The enrichment factors (EFs) were defined as the ratio of the analytes concentration in the extracted phase (Cexp ) to the initial concentration of analytes (C0 ) in the sample: EF = Cexp /C0 . Under the optimum extraction and derivatization conditions above, results of EFs were 170.5, 240.3, and 283.2 for BPA, OP, and NP, respectively.
3.6 Comparison with reported methods In (Table 2), LODs of the proposed method were compared with those of the reported methods for the determination of BPA, OP, and NP in different kinds of real samples [2, 9, 33–38]. Compared with pressurized liquid extraction [2], SPME [34], SDME [35], and SPE [9, 36], DLLME generally showed relatively lower LODs [33, 35]. Moreover, the proposed method in this study using EASC as derivatization reagents was also compared with those of other derivatization reagents or without a derivatization procedure. The LODs of these methods with a derivatization procedure by HPLC–FLD [36–38] were relatively lower than LC–MS/MS [2], GC–MS [9], and HPLC–UV [33–35] without derivatization. www.jss-journal.com
2762
J. Sep. Sci. 2014, 37, 2757–2763
T. Lv et al.
Table 1. Linear regression equations, correlation coefficients, LODs, and LOQs of three phenol derivatives
Phenol derivatives
BPA OP NP
Linear regression equations Y = AX + Ba)
Y = 2.78X – 1.78 Y = 1.07X + 0.44 Y = 1.03X + 3.06
R2
0.9999 0.9998 0.9998
Soft drinks
Dairy products
LOD (ng/L)
LOQ (ng/L)
LOD (g/kg)
LOQ (g/kg)
0.5 1.0 1.2
1.6 3.3 3.9
0.01 0.03 0.04
0.03 0.10 0.12
Detection method
LOD (g/kg)
a) X is concentration of BPA, OP, and NP (ng/L); Y is peak area. Table 2. Comparison of the proposed method with reported methods
Sample type
Extraction method
Derivatization reagent
Ref.
BPA
OP
NP
Milk powder Infant formula River and tap water Wastewater
Pressurized liquid extraction SPE DLLME SPME
– – – –
LC–MS/MS GC–MS HPLC–UV HPLC–UV
5.0 0.5 0.07 0.3
3.0 – – 0.8
5.0 – –
[2] [9] [33] [34]
Seawaters
SDME DLLME
– –
HPLC–UV HPLC–UV
4.0 0.7
9.0 0.8
9.0 1.6
[35]
Urine Sediment Plastic Milk and milk powder
SPE SPE SPE UA-DLLME
p-Nitrobenzoyl chloride CEOCa) DIB-Clb) EASC
HPLC–FLD HPLC–FLD HPLC–FLD HPLC–FLD
2.7 0.1 0.67 0.01
– 0.2 – 0.03
2.9 0.2 1.3 0.04
[36] [37] [38] This work
–, not included in the method. a) CEOC, 2-(9H-carbazol-9-yl)ethyl carbonochloridate. b) DIB-Cl, 4-(4,5-diphenyl-1H-imidazol-2-yl)benzoyl chloride.
However, LODs of this study for BPA, OP, and NP were about 7–900 times lower than the reported methods listed in (Table 2) [2, 9, 33–38]. This should be the benefit of the use of the hyphenated technique of UA-DLLME and derivatization followed by HPLC and fluorescence detection, which was not reported for the determination of BPA, OP, and NP. In conclusion, the proposed strategy of UA-DLLME combined with EASC derivatization could greatly increase the sensitivity and the selectivity, and the matrix effect was also effectively improved.
3.7 Application to real samples This method was successfully applied for the determination of BPA, OP, and NP in different soft drinks and dairy products. In this study, three kinds of different brands of cola, green tea, milk, and milk powder were used as real samples for analysis. Figure 2B and C shows the representative chromatograms of a green tea and a milk powder sample with their spiked ones. All the samples were analyzed in triplicate and the mean values were reported as the results (Supporting Information Table S2). For six kinds of soft drink sam C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ples, BPA was found with concentrations ranging from 5.3 to 78.9 ng/L; OP was found in a cola sample and a green tea sample with concentrations at 20.4 and 7.2 ng/L; NP was found in two cola samples with concentrations at 18.5 and 5.8 ng/L, respectively. For six kinds of dairy product samples, BPA was detectable in two milk powder samples and one milk sample, at levels of 0.8, 8.4, and 128.7 g/kg; OP and NP were found in two milk samples and one milk powder sample of the same at levels of 1.5, 0.3, and 29.5 g/kg and 13.7, 3.5, and 36.7 g/kg, respectively. According to the calculation method for the tolerable daily intake (TDI) of the European Food Safety Authority (EFSA) [2], the results in this study showed that BPA, OP, and NP were present in soft drinks and dairy products but levels were very low within the limits of the TDI.
4 Concluding remarks In this study, a highly sensitive and selective UA-DLLME pretreatment technique coupled with derivatization followed by HPLC–FLD was developed and validated for the determination of BPA, OP, and NP in soft drinks and dairy products. The combination of UA-DLLME and derivatization www.jss-journal.com
J. Sep. Sci. 2014, 37, 2757–2763
effectively improved the sensitivity, selectivity, and matrix effect of EDCs. This hyphenated technique is the first of its kind for the analysis of BPA, OP, and NP and it could be further extended for the analysis of other homologous analytes. The authors acknowledge the financial support from the National Natural Science Foundation of China (81303179, 21275089) and the Foundation of Qufu Normal University (2012019). The authors have declared no conflict of interest.
5 References
Sample Preparation
2763
[17] Schmitz-Afonso, I., Loyo-Rosales, J. E., de la Paz Aviles, M., Rattner, B. A., Rice, C. P., J. Chromatogr. A 2003, 1010, 25–35. [18] You, J., Zhao, H., Sun, Z., Xia, L., Yan, T., Suo, Y., Li, Y., J. Sep. Sci. 2009, 32, 1351–1362. [19] Zhang, S., You, J., Sun, Z., Song, C., Ning, S., Zhao, C., Suo, Y., Microchem. J. 2012, 103, 90–96. ´ ´ [20] Zafra, A., del Olmo, M., Suarez, B., Hontoria, E., Navalon, A., Vı́lchez, J., Water Res. 2003, 37, 735–742. [21] Vega-Morales, T., Sosa-Ferrera, Z., Santana-Rodr´ıguez, J., J. Hazard. Mater. 2010, 183, 701–711. [22] Basheer, C., Parthiban, A., Jayaraman, A., Lee, H. K., Valiyaveettil, S., J. Chromatogr. A 2005, 1087, 274–282.
[1] Diamanti-Kandarakis, E., Bourguignon, J. P., Giudice, L. C., Hauser, R., Prins, G. S., Soto, A. M., Zoeller, R. T., Gore, A. C., Endocr. Rev. 2009, 30, 293–342.
[23] Kawaguchi, M., Ito, R., Endo, N., Okanouchi, N., Sakui, N., Saito, K., Nakazawa, H., J. Chromatogr. A 2006, 1110, 1–5.
˜ [2] Ferrer, E., Santoni, E., Vittori, S., Font, G., Manes, J., Sagratini, G., Food Chem. 2011, 126, 360–367.
[24] Xu, L., Basheer, C., Lee, H. K., J. Chromatogr. A 2007, 1152, 184–192.
[3] Adamusova, H., Bosakova, Z., Coufal, P., Pacakova, V., J. Sep. Sci. 2014, 37, 885–905.
[25] Lin, C. Y., Fuh, M. R., Huang, S. D., J. Sep. Sci. 2011, 34, 428–435.
[4] Mezcua, M., Martinez-Uroz, M. A., Gomez-Ramos, M. M., Gomez, M. J., Navas, J. M., Fernandez-Alba, A. R., Talanta 2012, 100, 90–106.
[26] Rezaee, M., Assadi, Y., Milani Hosseini, M.-R., Aghaee, E., Ahmadi, F., Berijani, S., J. Chromatogr. A 2006, 1116, 1–9.
[5] Fernandes, A., Costley, C., Rose, M., Food Addit. Contam. 2003, 20, 846–852.
´ ´ ´ [27] Lopez-Darias, J., German-Hern andez, M., Pino, V., Afonso, A. M., Talanta 2010, 80, 1611–1618.
[6] Guenther, K., Heinke, V., Thiele, B., Kleist, E., Prast, H., Raecker, T., Environ. Sci. Technol. 2002, 36, 1676–1680.
[28] Vinas, P., Campillo, N., Lopez-Garcia, I., HernandezCordoba, M., Anal. Bioanal. Chem. 2014, 406, 2067–2099.
[7] Schecter, A., Malik, N., Haffner, D., Smith, S., Harris, T. R., Paepke, O., Birnbaum, L., Environ. Sci. Technol. 2010, 44, 9425–9430. [8] Yonekubo, J., Hayakawa, K., Sajiki, J., J. Agric. Food Chem. 2008, 56, 2041–2047. [9] Cao, X. L., Dufresne, G., Belisle, S., Clement, G., Falicki, M., Beraldin, F., Rulibikiye, A., J. Agric. Food. Chem. 2008, 56, 7919–7924. [10] Kawaguchi, M., Inoue, K., Yoshimura, M., Ito, R., Sakui, N., Okanouchi, N., Nakazawa, H., J. Chromatogr. B 2004, 805, 41–48. [11] Jiao, Y., Ding, L., Fu, S., Zhu, S., Li, H., Wang, L., Anal Methods 2012, 4, 291–298. ´ ˜ [12] Llorca-Porcel, J., Mart´ınez-Parreno, M., Mart´ınezSoriano, E., Valor, I., J. Chromatogr. A 2009, 1216, 5955– 5961.
[29] Saraji, M., Boroujeni, M. K., Anal. Bioanal. Chem. 2014, 406, 2027–2066. [30] Parrilla Vazquez, M. M., Parrilla Vazquez, P., Martinez Galera, M., Gil Garcia, M. D., Ucles, A., J. Chromatogr. A 2013, 1291, 19–26. [31] Campone, L., Piccinelli, A. L., Pagano, I., Carabetta, S., Di Sanzo, R., Russo, M., Rastrelli, L., J. Chromatogr. A 2014, 1334, 9–15. [32] Iparraguirre, A., Prieto, A., Navarro, P., Olivares, M., ´ Fernandez, L., Zuloaga, O., Anal. Bioanal. Chem. 2011, 401, 339–352. [33] Rezaee, M., Yamini, Y., Shariati, S., Esrafili, A., Shamsipur, M., J. Chromatogr. A 2009, 1216, 1511–1514. [34] Penalver, A., Pocurull, E., Borrull, F., Marce, R., J. Chromatogr. A 2002, 964, 153–160.
ˇ ıcek, ˇ ´ A., Vlcek, ˇ [13] Matej´ D., Grycova, J., J. Sep. Sci. 2013, 36, 1097–1103.
[35] Lopez-Darias, J., German-Hernandez, M., Pino, V., Afonso, A. M., Talanta 2010, 80, 1611–1618.
´ ´ [14] Ballesteros-Gomez, A., Rubio, S., Perez-Bendito, D., J. Chromatogr. A 2009, 1216, 449–469.
[36] Mao, L., Sun, C., Zhang, H., Li, Y., Wu, D., Anal. Chim. Acta 2004, 522, 241–246.
[15] Braunrath, R., Podlipna, D., Padlesak, S., Cichna-Markl, M., J. Agric. Food Chem. 2005, 53, 8911–8917.
[37] Zhang, S. J., You, J. M., Song, C. H., Chen, G., Suo, Y. R., Anal Methods 2012, 4, 4030–4036.
´ [16] Garc´ıa-Prieto, A., Lunar, M. L., Rubio, S., Perez-Bendito, D., Anal. Chim. Acta 2008, 630, 19–27.
[38] Sun, Y., Wada, M., Kuroda, N., Hirayama, K., Nakazawa, H., Nakashima, K., Anal. Sci. 2001, 17, 697–702.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
