2591
J. Sep. Sci. 2014, 37, 2591–2598
Linli Zhu Hui Xu Key Laboratory of Pesticide and Chemical Biology, Ministry of Education, Institute of Environmental Chemistry, College of Chemistry, Central China Normal University, Wuhan, China Received April 7, 2014 Revised June 19, 2014 Accepted June 22, 2014
Research Article
Magnetic graphene oxide as adsorbent for the determination of polycyclic aromatic hydrocarbon metabolites in human urine Detection of monohydroxy polycyclic aromatic hydrocarbons metabolites in urine is an advisable and valid method to assess human environmental exposure to polycyclic aromatic hydrocarbons. In this work, novel Fe3 O4 /graphene oxide composites were prepared and their application in the magnetic solid-phase extraction of monohydroxy polycyclic aromatic hydrocarbons in urine was investigated by coupling with liquid chromatography and mass spectrometry. In the hybrid material, superparamagnetic Fe3 O4 nanoparticles provide fast separation to simplify the analytical process and graphene oxide provides a large functional surface for the adsorption. The prepared magnetic nanocomposites were characterized by X-ray diffraction, scanning electron microscopy, transmission electron microscopy, and vibrating sample magnetometry. The experimental conditions were optimized systematically. Under the optimal conditions, the recoveries of these compounds were in the range of 98.3–125.2%, the relative standard deviations ranged between 6.8 and 15.5%, and the limits of detection were in the range of 0.01–0.15 ng/mL. The simple, quick, and affordable method was successfully used in the analysis of human urinary monohydroxy polycyclic aromatic hydrocarbons in two different cities. The results indicated that the monohydroxy polycyclic aromatic hydrocarbons level in human urine can provide useful information for environmental exposure to polycyclic aromatic hydrocarbons. Keywords: Graphene oxide / Liquid chromatography with mass spectrometry / Magnetic solid-phase extraction / Monohydroxy polycyclic aromatic hydrocarbons / Urine DOI 10.1002/jssc.201400363
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Polycyclic aromatic hydrocarbons (PAHs) are a common carcinogen in mammals, which are mainly from incomplete combustion of organic matrices. Exposure to these environmental contaminants and their potential damage to public health has become a worldwide concern [1–3]. Once PAHs in the polluted air are absorbed into human body, they can be converted to epoxide intermediates, such as diol-epoxide, which is a highly reactive metabolite endangering cellular DNA [4]. These intermediates are the foundation of carcinogenic substances and the potential risk to human health is a chronic and accumulated process. It has been proved that exposure to PAHs may disrupt sperm DNA [5–8] and interCorrespondence: Dr. Hui Xu, Luoyu Road 152, Wuhan, Hubei Province, China E-mail: [email protected] Fax: +86-27-67867955
Abbreviations: EF, enrichment factor; MSPE, magnetic solidphase extraction; MNP, magnetic nanoparticle; PAH, polycyclic aromatic hydrocarbon C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
fere with human male fertility [9]. It has also been demonstrated that exposure to fumes and aerosols of bitumen may contribute to increased DNA strand breakage in white blood cells [10]. At present, urinary monitoring has become an increasingly relevant tool for evaluating human exposure to chemicals and pollutants [11–14], because it has many advantages, such as no invasion, pervasiveness, and accessibility. Analytical capabilities are at the core of biomonitoring for contaminants in complex urine matrices. Many analytical methods have been developed for the measurement of urinary metabolites. However, several main problems still challenge the existing analytical method of human biomonitoring. The low concentrations of metabolites in urine might be beyond instrumental LODs, and the coexisting compounds (organic solutes, inorganic ions, and macromolecules) in urine might lead to matrix interference. Therefore, an appropriate sample preconcentration step before chromatographic and mass spectrometric analysis is definitely necessary. Conventional preconcentration methods include LLE [15] and SPE [16]. Colour Online: See the article online to view Figs. 1–4 in colour. www.jss-journal.com
2592
J. Sep. Sci. 2014, 37, 2591–2598
L. Zhu and H. Xu
However, these two methods suffer from disadvantages as they are labor intensive, time consuming, and solvent consuming. SPME, a solvent-free extraction procedure, was introduced by Pawliszyn in 1990 [17]. It is a simple, fast, and green sample preparation technique, but the fiber is fragile and expensive. A new magnetic solid-phase extraction method (MSPE) based on magnetic nanoparticles (MNPs) was developed and had received increasing interest in recent years [18,19]. MNPs have high extraction efficiency because of the high surface area-to-volume ratio compared with other adsorbents. They exhibit the advantage of rapid separation from matrix due to the excellent magnetic response. MSPE has been widely applied to analysis of drugs [20, 21], environmental water samples [22,23], and biological samples [24,25]. Besides the superior magnetic response, specific surface modification, high binding capacity, and extraction efficiency are also vital for magnetic adsorbents. Therefore, many studies are focused on the surface modifications of MNPs [26–29]. For example, alkyl silane modified MNPs (C8 -amine-functionalized MNPs) were prepared and used to determine aldehyde in human urine [26]. Dodecylbenzene sulfonate coated magnetite nanoparticles were used as an adsorbent for the determination of trace amounts of ammonium in water sample [27]. Water-dispersible, photocatalytic Fe3 O4 @TiO2 coreshell magnetic nanoparticles were prepared for capturing and photocatalytically destroying some endocrine-disrupting chemicals [28]. A recent review about the application of functionalized magnetic nanoparticles in sample preparation has been reported by Xie [29]. Graphene (G) and graphene oxide (GO) are emerging and promising carbon-based nanomaterials, they have sparked great interest of scientists due to their unique 2D planar monolayer structure, superior mechanical strength, remarkable thermal and chemical stability, and ultrahigh specific surface area. For GO, the abundant oxygen-containing hydrophilic functional groups on the surface and edges can provide specific adsorption sites and stable dispersion in water. These properties make GO very attractive for various applications including sample preparation and a relevant review has been reported recently [30]. However, the direct use of GO in powder form may be inconvenient in handling and cleanup, leakage and blocking may occur when nanosized GO is used as a sorbent in SPE. Magnetic graphene-based MSPE can solve these problems effectively. Moreover, the recovery of nanoparticles from suspension can be realized easily by a magnet. Novel materials have been prepared and applied for the analysis of neonicotinoid pesticides from pear and tomato samples [31], phthalate esters from water and beverages [32] and soybean milk samples [33]. In this paper, magnetic graphene oxide composite was prepared and applied for the determination of PAH metabolites in human urine. Three polycyclic aromatic hydrocarbons, 2-naphthol, 3-hydroxy-phenanthrene, 1-hydroxypyrene were selected as model compounds. The feasibility of Fe3 O4 /GO nanocomposite as MSPE adsorbent in biomonitoring was estimated. Based on the MNPs, an MSPE–LC–MS C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
method was established. The influence of sorbent amount, extraction conditions, and desorption conditions were optimized. The developed method was applied to the analysis of human urinary OH-PAHs in two different cities of Hubei Province, China.
2 Materials and methods 2.1 Apparatus An Alliance e2695 separations module, as the HPLC system (Waters, Manchester, UK) and Massllynx 4.1 software were used to achieve the separation of the three PAH metabolites. HPLC analysis was obtained by a Waters symmetry C8 column [150 mm × 2.1 mm (id), 3.5 m] at the flow rate of 0.2 mL/min and the column temperature was 30⬚C. To get the best chromatographic separation, ultrapure water containing 5 mM/L ammonium acetate (A) and acetonitrile (B) were selected as a gradient LC system. The linear gradient elution was as follows: initial state was 50% of A, and then linearly increased to 100% within 4 min; and got back to 50% in 4 min, and equilibrated for 3 min. Injection volume was 10 L. The MS system was an Acquity TQD instrument (Waters, Manchester, UK), which was equipped with a Z-spray source for negative ESI. Selective ion reaction was executed through MS mode. Some other MS parameters were set as follows, capillary voltage: 2.5 kV, extraction voltage: 2 V, ion source temperature: 120⬚C, solvent temperature: 350⬚C, solvent gas flow: 550 L/h. The confirmation and quantification ions (m/z) were 143.09, 193.11, and 217.11 for 2-naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene, respectively.
2.2 Reagents 2-Naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene were obtained from Dr. Ehrenstorfer (Germany). HPLCgrade ethanol was purchased from Tedia (USA). HPLCgrade methanol and acetonitrile were obtained from Merck (Darmstadt, Germany). Ultrapure water was used in all experiments (Millipore Simplicity 185, USA). Glucuronidase/arylsulfatase was obtained from Merck (Darmstadt, Germany). Iron(III) chloride hexahydrate (FeCl3 ·6H2 O), iron chloride tetrahydrate (FeCl2 ·4H2 O), nitric acid (HNO3 ), hydrogen peroxide (H2 O2 ), and potassium permanganate (KMnO4 ) were all of analytical grade and purchased from Shanghai chemical reagents company (Shanghai, China). Graphite powder (325 mesh, 99.95%) was purchased from Jinrilai graphite (Qingdao, China). Other chemical reagents were all of analytical grade.
2.3 Preparation of adsorbent material Fe3 O4 nanoparticles were generated with a modified Massart method (for details see Supporting Information S2) [34]. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2591–2598
Expanded graphite was used to create graphite oxide according to a modified Hummer’s method (for details see Supporting Information S2) [35]. The synthesis of Fe3 O4 /GO nanocomposite is illustrated in Fig. 1. First, the prepared Fe3 O4 nanoparticles were dispersed in 1 mol/L HNO3 to obtain a positive surface charge. Second, graphite oxide was exfoliated to graphene oxide by ultrasound for an hour. Last, the modified Fe3 O4 and GO were mixed and mechanically agitated for 4 h to generate Fe3 O4 /GO composites [36].
Sample Preparation
2593
tration of 5 g/L. After vortex dispersion, 500 L sorbent was added in the 2 mL standard solution. Second, the mixed suspension was agitated thoroughly on a vortex mixer for 3 min to reach adsorption equilibrium. Immediately, the Fe3 O4 /GO was isolated from the sample solution with a magnet. Third, 50 L ethanol and 50 L acetonitrile were used to elute the target analytes from the sorbent (3 min). Finally, an aliquot of 10 L filtered supernatant was injected for LC–MS analysis.
3 Results and discussion 2.4 Characterization of synthetic materials
3.1 Characterization of Fe3 O4 /GO nanocomposites
Powder X-ray diffraction patterns were obtained on a Bruker D8 Advance X-ray diffractometer with CuK␣ radiation (k = 1.54178 Å). SEM images were obtained on a LEO 1450VP scanning electron microscope. Zeta potential measurements were performed using a nanoparticle zeta potential analyzer (Zen3690, Malvern, UK), and the graphene oxide and Fe3 O4 samples were diluted to 1 mg/mL before measurements. The specific surface area was obtained on high-precision BELSORP-mini specific surface area and aperture determinator (BEL JAPAN). Magnetic properties of the materials were characterized by a PPMS-9 vibrating sample magnetometer (QUANTOM, USA).
The morphology and structure of the prepared Fe3 O4 nanoparticles, graphene oxide, and Fe3 O4 /GO nanocomposites were examined by SEM and TEM (Fig. 2A–C). From Fig. 2A and A1, we can see that Fe3 O4 nanoparticles with the average diameter of about 20–30 nm were aggregated together. For GO, a planar monolayer structure with the smooth, wrinkle-like and silk-like morphology was observed (Fig. 2B and B1). SEM and TEM images of the hybrid composite confirmed the attachment of Fe3 O4 nanoparticles onto the GO surface (Fig. 2C and C1). XRD was used to characterize the phase structure of Fe3 O4 , graphite oxide, and Fe3 O4 /GO (Fig. 2D–F). It was found that the Fe3 O4 phase (JCPDS file No. 26–1136) was synthesized. The position of the diffraction peaks at 2 = 36.74, 58.16, and 64.19⬚ can be assigned to (311), (511), and (440) reflections, respectively (Fig. 2D). The graphite oxide (Fig. 2E) presented a very sharp diffraction peak at 2 = 9.29⬚, which indicated that the graphite oxide was successfully synthesized. The diffraction peak at 2 = 9.29⬚ can be assigned to the (002) reflection. In Fig. 2F, the characteristic peaks of Fe3 O4 remained unchanged, but the diffraction peak for graphene oxide increased to 2 = 30.11⬚, which indicates the conversion from graphite oxide to graphene oxide after ultrasonic treatment, a similar change has been observed previously [36]. In addition, the surface electric charges of Fe3 O4 and the GO were also measured, the result showed that the zeta potentials of acidulated Fe3 O4 and GO were 11.9 and −29.6 mV, respectively. Fe3 O4 nanoparticles dispersed in 1 mol/L HNO3 possessed positive surface charge. The surface of the as prepared GO sheets were highly negatively charged, which was apparently a result of ionization of the carboxylic acid and phenolic hydroxyl groups on the surface [36]. The data suggest that electrostatic interaction might play an important role in their combination. These data presented above indicated that the Fe3 O4 /GO composites were successfully prepared based on a simple one-step surface charge-driven self-assembly. The magnetic properties of Fe3 O4 and Fe3 O4 /GO were also determined by vibrating sample magnetometry. As shown in Supporting Information Fig. S2, we can see that the composites were superparamagnetic. The specific saturation magnetization of Fe3 O4 , Fe3 O4 /GO (10:1 m/m) and Fe3 O4 /GO (5:1 m/m) was 65.98, 58.78, and 40.53 emu/g, respectively. The surface area of synthetic Fe3 O4 /GO nanocomposites tested by nitrogen
2.5 Preparation of standard solution Two types of standard solution were prepared. Stock solutions of 1000 mg/L 2-naphthol, 500 mg/L 3-hydroxyphenanthrane, and 1000 mg/L 1-hydroxypyrene were prepared separately in methanol, stored at −20⬚C. The daily standard working solution was prepared by dilution and mixing the stock solution according to demand. The calibration solution used in method validation was prepared by spiking the standard solution into real urine matrix (the signals of real urine were subtracted from those of the spiked urine).
2.6 Pretreatment of urine samples Thirty eight urine samples were collected in the morning from nonsmoking men aged 18–22. Before use, urine samples were filtered with a 0.22 m filter membrane. Ten microliters of -glucuronidase/arylsulfatase and a certain amount of standard solution were added to 10 mL of the urine pool (pH 5.5) in a 12 mL glass vial. Then, the mixture was incubated at 37⬚C for 4 h away from light.
2.7 Magnetic solid-phase extraction procedure The MSPE process is as follows (Supporting Information Fig. S1). First, Fe3 O4 /GO nanocomposites were dispersed into deionized water to create a suspension with the concen C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
2594
L. Zhu and H. Xu
J. Sep. Sci. 2014, 37, 2591–2598
Figure 1. Schematic illustration of the fabrication of Fe3 O4 /GO nanocomposite.
adsorption by the BET method was 109.58 m2 /g. The large surface area of composites provided adequate adsorption sites for the following extraction application.
3.2 Selection of extraction conditions The mass ratio between Fe3 O4 and GO was a crucial factor in the extraction. Because iron content has a great impact on the magnetism of synthetic materials. A higher Fe3 O4 percent-
age means more rapid magnetic separation. As shown in the Supporting Information Fig. S2, Fe3 O4 /GO (10:1 m/m) has a stronger superparamagnetism than Fe3 O4 /GO (5:1 m/m), the results indicated that increased loading amounts of Fe3 O4 can ensure the stronger superparamagnetism of nanocomposites. While for GO, a higher ratio means a larger surface area, more adsorption sites, and better dispersion in aqueous solution. In order to investigate the adsorptive capabilities of Fe3 O4 /GO composites with different proportions for PAHs, five kinds of nanocomposites (1:1, 5:1, 10:1, 15:1, 20:1,
Figure 2. SEM images (A, B, C) and XRD patterns (D, E, F) of the freshly prepared nanoparticles. Insets are TEM images (A-1, B-1, C-1). (A) and (D): Fe3 O4 , (B): GO, (C) and (F): Fe3 O4 /GO, (E): Graphite oxide. Scale bars for A, B, C, A-1, B-1, and C-1 are 10 nm, 1 m, 100 nm, 50 nm, 1 m, and 100 nm, respectively.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2591–2598
Sample Preparation
2595
Figure 3. Effect of the extraction time (A) and desorption time (B) on the peak areas. Conditions: 2 mL sample solution, concentration of 2-naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene: 25, 12.5, and 25 ng/mL, Fe3 O4 /GO absorbent: 500 L, 2 g/L.
500 L, 2 g/L) were added into 2 mL aqueous solution containing 25 ng/mL of 2-naphthol, 12.5 ng/mL of 3-hydroxyphenanthrene and 25 ng/mL of 1-hydroxypyrene. The best result was obtained at the ratio of 10:1 (Fe3 O4 /GO, m/m, for details see Supporting Information Fig. S3a). Meanwhile, the extraction efficiency of bare Fe3 O4 was also investigated. The result showed that bare Fe3 O4 can also adsorb OH-PAHs, while the extraction efficiency was far less than those of Fe3 O4 /GO. Thus, it can be concluded that the adsorption of OH-PAHs is ascribed mainly to GO. In addition, the effect of the Fe3 O4 /GO amount on the extraction efficiency was further investigated. Nanocomposites at the concentration of 5 g/L showed the best extraction efficiency (Supporting Information Fig. S3b). Higher concentration means more adsorbents, which is beneficial for extraction, but it is inconvenient for the withdrawing of desorption solvent in desorption procedure. Accordingly, 5 g/L of 10:1 Fe3 O4 /GO was used as adsorbent in the following experiments. The mixture model of absorption and desorption was also investigated. The result showed that vortex was the best way for this work, because vortex can accelerate the mass transfer process of analytes from sample solution to dispersive material. Furthermore, the effect of adsorption time was studied in the range of 1 to 6 min. As the extraction time profile shows in Fig. 3A, there was no significant difference in the peak area of 2-naphthol and 1-hydroxypyrene with time. For 3-hydroxy-phenanthrene, the best extraction efficiency was obtained at 4 min. The results indicated that 4 min was adequate to perform the extraction. The rapid extraction time was attributed to high surface area provided by the nanosized magnetic sorbents. Thus, we chose 4 min as the optimum extraction time.
3.3 Selection of desorption conditions An effective elution can ensure high recovery of extraction method. In our work, five different organic solvents C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(methanol, ethanol, acetonitrile, methanol/acetonitrile, and ethanol/acetonitrile) were selected as eluents and their desorption efficiencies were investigated. The results showed that the mixture of acetonitrile and ethanol had a higher desorption capacity than other solvents (Supporting Information Fig. S4a). And, the ratio of acetonitrile to ethanol (2:8, 3:7, 5:5, 7:3, 8:2, v/v) was studied further, the results showed that the mixed solvent at the ratio of 1:1 offered better desorption efficiency (Supporting Information Fig. S4b). The desorption time profile was also investigated, as can be seen from Fig. 3B, the maximum peak signal was obtained at 3 min for 2-naphthol and 3-hydroxyphenanthrene. While for 1-hydroxypyrene, the signal decreased continuously with time. Hence, 3 min was selected as the appropriate desorption time. The eluent volume was also optimized, 100 L was selected to ensure high enrichment and recovery. In summary, analytes were vortex extracted for 4 min with 5 g/L of Fe3 O4 /GO (10:1, m/m) nanocomposites as adsorbent and desorbed in 100 L acetonitrile and ethanol mixture (1:1, v/v) for 3 min in the following investigation. Under optimal conditions, the enrichment factors of three target compounds were calculated. Enrichment factor (EF) is defined as the ratio of the peak area of an extracted analyte to that in the original sample determined by direct injection of the solution. The data for 2-naphthol, 3hydroxyphenanthrene, and 1-hydroxypyrene were 5.4, 17.1, and 19.5, respectively (theoretical enrichment factor: 20). LogP (octanol/water partition coefficient, an indicator for hydrophobicity) values were 2.71, 3.909, and 4.43 for three analytes. The Fe3 O4 /GO composites exhibited high EF values for 3-hydroxy-phenanthrene (17.1) and 1-hydroxypyrene (19.5), suggesting the high extraction selectivity toward hydrophobic compounds, the similar results were also reported in previous report [37]. It is well known that there exists the – stacking interaction between GO and the benzene ring of PAHs, and the interaction increases with the ring number of PAHs [38–40]. It was reported that the graphene oxide is more likely to adsorb the targets with more benzene rings. This order of EFs was also consistent with the ring number www.jss-journal.com
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
103.6 125.2 99.7 12.6 9.8 11.1 6.8 9.9 9.8 15.5 9.3 8.2 1–3000 0.2–300 0.1–300 y = 208.16x + 4715.9 y = 5072.4x + 28034 y = 10461x + 83796 2-Naphtholb) 3-Hydroxy-phenanthrene 1-Hydroxypyrene
a) The concentrations unit of analytes in RSD and recovery experiments was ng/mL. b) The low, medium, high concentrations of 2-naphthol were 1, 10, and 100 ng/mL in RSD experiments, respectively.
13.4 14.0 10.2
10 1
12.0 9.6 6.8 7.1 13.1 8.5 0.15 0.13 0.01
0.1 0.1
1
Inter-day RSDa) (%, n=5) Intra-day RSDa) (%, n = 5) Linear range Regression equation
To further demonstrate the advantages of the proposed method, the proposed method was compared with previous published reports, which used three conventional sample preparation methods (LLE [15], SPE [16], SPME [17]). It is clearly seen from Table 2 that our method showed obvious advantages over the other three methods in analysis time and recovery. With nanocomposites as sorbent, fast extraction can be obtained due to the high surface area-to-volume ratio of nanocomposites. High recovery data (98.3–125.2%) indicated that the satisfactory accuracy of our method. Separation of magnetic sorbents from large-volume samples is convenient and fast with the aid of magnet and no filtration and centrifugation is needed. Sensitivity of our method (LC– MS) was a little lower than that of a GC–MS method, the possible explanation is the ionization of OH-PAHs in electrospray ion sources (LC–MS) is poorer than in electron ionization source (GC–MS) due to the weak polarity of analytes. In summary, this method possesses many advantages, including rapid analysis, easy operation, satisfactory accuracy, and adequate sensitivity.
Analyte (ng/mL)
3.5 Method comparison
Table 1. Method validation data for the established MSPE-LC-MS method
r
LOD
A series of quantitative experiments in regard to linear range, correlation coefficient, and LOD, reproducibility, and recovery were conducted to examine the proposed method in urine matrix under the optimal experimental conditions. Linear regression analysis was presented using peak area against concentration of analytes. Table 1 lists the quantitative parameters of the proposed MSPE method. The LOD, which was calculated using S/N = 3, was 0.01–0.15 ng/mL. Calibration curves of three analytes exhibited good linearity with a correlation coefficient above 0.9945. The precision was investigated with five parallel experiment results of within-day precision and between-day at three concentration levels (0.1, 1, 10 ng/mL). The results in Table 1 show that RSD values ranged from 6.8 to 14.0%. Recovery experiments of MSPE procedure were performed by measuring the known concentrations analytes in urine samples of healthy people at two spiked concentrations (10, 100 ng/mL). The average recoveries were in range of 98.3–125.2% (Table 1). Satisfactory recovery demonstrated that the real urine matrix has little interference on the performance of MSPE and it could be concluded that the MSPE–LC–MS analytical method is reliable [41, 42].
10
3.4 Method validation
0.9959 0.9959 0.9945
100
Recoverya) (%)
of OH-PAHs. Based on the result, we can deduce that both – stacking interaction and hydrophobic interaction play important roles in the extraction of OH-PAHs.
115.4 98.3 117.3
J. Sep. Sci. 2014, 37, 2591–2598
L. Zhu and H. Xu
10
2596
www.jss-journal.com
Sample Preparation
J. Sep. Sci. 2014, 37, 2591–2598
2597
Table 2. Method comparison
Method
Extraction timec) (min)
LOD (ng/mL)
Recovery (%)
RSD (%)
Organic solvent
References
LLE(GC–IDHRMS)a)
>80 min
0.007–0.012
46–72
2.9–11
[15]
SPE(LC–EI-MS)
—
10–100
—
3.7–8.8
SPME(ID-GC–MS) MSPE(HPLC–MS)
>165 min 8 min
0.00078–0.0158 0.01–0.15
6–47 98.3–125
2.4–18.7 6.8–15.5
10 mL pentane; 5 L dodecane; 20 L toluene; 10 L MSTFAb) 3 mL methanol 6 mL fomic acid 15 L MSTFA 100 L ethanol and acetonitrile
[16] [17] This work.
a) Isotope dilution high-resolution MS. b) N-methyl-N-(trimethylsilyl)trifluoroacetamide. c) The time of treatment of urine is not included.
3.6 Application in real samples This proposed method was applied for the determination of three OH-PAHs in the urine of nonsmoking men aged 18–20 (n = 38) who lived in two cities with different environments (Wuhan and Enshi). The preliminary research aimed to evaluate the possibility of the method for biological monitoring of environment exposure to PAHs. With a current population exceeding nine million, the city of Wuhan is an emerging mega-city in Central China. It is also one of the most important industrial cities. And it has been reported that the air in Wuhan is highly polluted [43]. PAHs, from emission of coal-burning power plants, steel plants, or automobile exhaust are the main atmospheric pollutants in Wuhan. Enshi, a famous ecological tourist city in Central China, is surrounded by forest-covered mountains and rivers. Agriculture and tourism are the main economical drivers of the city. Wuhan and Enshi are ideal sites for the comparative study of environment pollution episodes, representative of Chinese cities with different populations and industrialization. Figure 4 depicts typical LC–MS–SIR chromatograms of blank and spiked samples of healthy volunteer. The results showed that the concentration of 1-hydroxypyrene (widely used as a kind of suitable PAH biomarker [44]) was higher than the other two analytes. Furthermore, there was a clear regional difference about the urinary concentration of 1hydroxypyrene in two cities. The average concentration of 1hydroxypyrene in place A (Wuhan, Hubei, China) was about 0.21 ng/mL and the detection rate was 84%. Contrastively, that concentration in place B (Enshi, Hubei, China) was much lower (0.12 ng/mL, the detection rate was only 68%), and some analytes cannot be detected using the proposed method. The difference in urinary concentration may be ascribed to the different exposure environment of the two cities. Compared with people in tourist city, distinctly higher urinary OHPAHs level was found in individuals who lived in industrial city. The results indicated that the urinary level of OH-PAHs could reflect the pollution level of people’s residence and the
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. LC–MS-SIR Chromatograms of OH-PAHs in blank urine sample (A); standard solution (B, 1 ng/mL); spiked urine (C, 10 ng/mL; D, 100 ng/mL); the insert is the urinary concentration of 1-hydroxypyrene in two cities (A: Wuhan, B: Enshi).
proposed method has the potential to determine pollutants and their metabolites in human urine.
4 Conclusion A facile determination of PAH metabolites in human urine based on the magnetic graphene oxide was reported in this paper. The hybrid Fe3 O4 /GO nanocomposites exhibit excellent magnetic response and satisfactory extraction performance toward three OH-PAHs. For assay of target analytes, the proposed method possesses numerous advantages, including cost effectiveness, time saving, ease of operation, and procedural simplicity. In the urine matrix, reasonable recovery, acceptable reproducibility, and low LOD were achieved for the method. In summary, this MSPE–HPLC–MS method has potential for the biomonitoring of OH-PAHs in human
www.jss-journal.com
2598
L. Zhu and H. Xu
J. Sep. Sci. 2014, 37, 2591–2598
urine, and it can provide useful information for environmental exposure to PAHs.
[17] Smith, C. J., Walcott, C. J., Huang, W., Maggio, V., Grainger, J., Patterson, D. G. Jr, J. Chromatogr. B 2002, 778, 157–164.
This work was supported by the National Science Foundation of China (Nos. 21175052 and 21105035), and the selfdetermined research funds of the CCNU from the colleges’ basic research and operation of MOE (No. CCNU14A05010).
[18] Robinson, P. J., Dunnill, P., Lilly, M. D., Biotechnol. Bioeng. 1973, 15, 603–606. ˇ ˇ ´ M., Safaˇ [19] Safaˇ rı́kova, rı́k, I., J. Magn. Magn. Mater. 1999, 194, 108–112.
The authors have declared no conflict of interest.
[20] Wang, Y., Wang, Y., Chen, L., Wan, Q.-H., J. Magn. Magn. Mater. 2012, 324, 410–417. [21] Gao, Q., Lin, C.-Y., Luo, D., Suo, L.-L., Chen, J.-L., Feng, Y.-Q., J. Sep. Sci. 2011, 34, 3083–3091.
5 References
[22] Liu, X., Lu, X., Huang, Y., Liu, C., Zhao, S., Talanta 2014, 119, 341–347.
[1] Grimmer, G., Jacob, J., Dettbarn, G., Naujack, K. W., Int. Arch. Occ. Env. Hea. 1997, 69, 231–239.
[23] Su, S., Chen, B., He, M., Hu, B., Xiao, Z., Talanta 2014, 119, 458–466.
[2] Masuck, I., Hutzler, C., Luch, A., J. Sep. Sci. 2011, 34, 2686–2696.
[24] Zheng, H. B., Mo, J. Z., Zhang, Y., Gao, Q., Ding, J., Yu, Q. W., Feng, Y. Q., J Chromatogr. A. 2014, 1329, 17–23.
[3] Marais, A. A. S., Laurens, J. B., J. Sep. Sci. 2005, 28, 2526–2533.
[25] Moazzen, M., Ahmadkhaniha, R., Gorji, M. E. h., Yunesian, M., Rastkari, N., Talanta 2013, 115, 957–965.
[4] Shimada, T., Fujii-Kuriyama, Y., Cancer Sci. 2004, 95, 1–6.
[26] Song, D., Gu, Y., Liang, L., Ai, Z., Zhang, L., Xu, H., Anal. Methods 2011, 3, 1418–1423.
´ [5] Sanchez-Guerra, M., Pelallo-Mart´ınez, N., D´ıaz-Barriga, ´ F., Rothenberg, S. J., Hernandez-Cadena, L., Faugeron, ´ S., Oropeza-Hernandez, L. F., Guaderrama-D´ıaz, M., Quintanilla-Vega, B., Mutat. Res.-Gen. Tox. En. 2012, 742, 66–71. [6] Gurbani, D., Bharti, S. K., Kumar, A., Pandey, A. K., Ana, G. R. E. E., Verma, A., Khan, A. H., Patel, D. K., Mudiam, M. K. R., Jain, S. K., Roy, R., Dhawan, A., Int. J. Hyg. Envir. Heal. 2013, 216, 553–565. ´ [7] Cebulska-Wasilewska, A., Pawłyk, I., Panek, A., Wiechec, A., Kalina, I., Popov, T., Georgieva, T., Farmer, P. B., Mutat. Res.-Fund. Mol. M. 2007, 620, 145–154. [8] Singh, R., Sram, R. J., Binkova, B., Kalina, I., Popov, T. A., Georgieva, T., Garte, S., Taioli, E., Farmer, P. B., Mutat. Res.-Fund. Mol. M. 2007, 620, 83–92. [9] Han, X., Zhou, N. Y., Cui, Z. H., Ma, M. F., Li, L. B., Cai, M., Li, Y. F., Lin, H., Li, Y., Ao, L., Liu, J. Y., Cao, J., Environ. Health Persp. 2011, 119, 652–657. [10] Marczynski, B., Raulf-Heimsoth, M., Preuss, R., Kappler, M., Kafferlein, H. U., Schott, K., Pesch, B., Mensing, T., Spickenheuer, A., Angerer, J., Bruning, T., Toxicol. Lett. 2006, 164, S282–S282. [11] Yusa, V., Ye, X., Calafat, A. M., TrAC, Trends Anal. Chem. 2012, 38, 129–142. [12] Guo, Y., Senthilkumar, K., Alomirah, H., Moon, H. B., Minh, T. B., Mohd, M. A., Nakata, H., Kannan, K., Environ. Sci. Technol. 2013, 47, 2932–2938. [13] Feldt, T., Fobil, J. N., Wittsiepe, J., Wilhelm, M., Till, H., ¨ Zoufaly, A., Burchard, G., Goen, T., Sci. Total Environ. 2014, 466–467, 369–376. [14] del Carmen Hurtado-Sanchez, ´ M., Espinosa-Mansilla, ´ ´ A., Rodr´ıguez-Caceres, M. I., Mart´ın-Tornero, E., Duran´ I., J. Sep. Sci. 2012, 35, 2575–2584. Meras, [15] Li, Z., Romanoff, L. C., Trinidad, D. A., Hussain, N., Jones, ¨ R. S., Porter, E. N., Patterson, D. G., Sjodin, A., Anal. Chem. 2006, 78, 5744–5751. [16] Andreoli, R., Manini, P., Bergamaschi, E., Mutti, A., Franchini, I., Niessen, W. M. A., J. Chromatogr. A. 1999, 847, 9–17. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[27] Eskandari, H., Shariati, M. R., Anal. Chim. Acta 2011, 704, 146–153. [28] Chalasani, R., Vasudevan, S., ACS Nano. 2013, 7, 4093–4104. [29] Xie, L., Jiang, R., Zhu, F., Liu, H., Ouyang, G., Anal. Bioanal. Chem. 2014, 406, 377–399. [30] Liu, Q., Shi, J., Jiang, G., TrAC, Trends Anal. Chem. 2012, 37, 1–11. [31] Ma, X., Wang, J., Sun, M., Wang, W., Wu, Q., Wang, C., Wang, Z., Anal. Methods 2013, 5, 2809–2815. [32] Wu, Q., Liu, M., Ma, X., Wang, W., Wang, C., Zang, X., Wang, Z., Microchim. Acta. 2012, 177, 23–30. [33] Wang, W., Ma, R., Wu, Q., Wang, C., Wang, Z., Talanta 2013, 109, 133–140. [34] Massart, R., IEEE T. Magn. 1981, 17, 1247–1248. [35] Hummers, W. S., Offeman, R. E., J. Am. Chem. Soc. 1958, 80, 1339–1339. [36] Han, Q., Wang, Z., Xia, J., Chen, S., Zhang, X., Ding, M., Talanta 2012, 101, 388–395. [37] Fan, J., Dong, Z., Qi, M., Fu, R., Qu, L., J Chromatogr. A 2013, 1320, 27–32. [38] Zhao, G., Jiang, L., He, Y., Li, J., Dong, H., Wang, X., Hu, W., Adv. Mater. 2011, 23, 3959–3963. [39] Zhao, G., Li, J., Wang, X., Chem. Eng. J. 2011, 173, 185– 190. [40] Wu, T., Cai, X., Tan, S., Li, H., Liu, J., Yang, W., Chem. Eng. J. 2011, 173, 144–149. [41] Yan, S., Qi, T.-T., Chen, D.-W., Li, Z., Li, X.-J., Pan, S.-Y., J Chromatogr. A 2014, 1347, 30–38. [42] Farajzadeh, M. A., Mogaddam, M. R. A., Ghorbanpour, H., J Chromatogr. A 2014, 1347, 8–16. [43] Qian, Z., He, Q., Lin, H.-M., Kong, L., Liao, D., Dan, J., Bentley, C. M., Wang, B., Environ. Res. 2007, 105, 380–389. [44] Jacob, J., Seidel, A., J. Chromatogr. B. 2002, 778, 31–47.
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2591–2598
Linli Zhu Hui Xu Key Laboratory of Pesticide and Chemical Biology, Ministry of Education, Institute of Environmental Chemistry, College of Chemistry, Central China Normal University, Wuhan, China Received April 7, 2014 Revised June 19, 2014 Accepted June 22, 2014
Research Article
Magnetic graphene oxide as adsorbent for the determination of polycyclic aromatic hydrocarbon metabolites in human urine Detection of monohydroxy polycyclic aromatic hydrocarbons metabolites in urine is an advisable and valid method to assess human environmental exposure to polycyclic aromatic hydrocarbons. In this work, novel Fe3 O4 /graphene oxide composites were prepared and their application in the magnetic solid-phase extraction of monohydroxy polycyclic aromatic hydrocarbons in urine was investigated by coupling with liquid chromatography and mass spectrometry. In the hybrid material, superparamagnetic Fe3 O4 nanoparticles provide fast separation to simplify the analytical process and graphene oxide provides a large functional surface for the adsorption. The prepared magnetic nanocomposites were characterized by X-ray diffraction, scanning electron microscopy, transmission electron microscopy, and vibrating sample magnetometry. The experimental conditions were optimized systematically. Under the optimal conditions, the recoveries of these compounds were in the range of 98.3–125.2%, the relative standard deviations ranged between 6.8 and 15.5%, and the limits of detection were in the range of 0.01–0.15 ng/mL. The simple, quick, and affordable method was successfully used in the analysis of human urinary monohydroxy polycyclic aromatic hydrocarbons in two different cities. The results indicated that the monohydroxy polycyclic aromatic hydrocarbons level in human urine can provide useful information for environmental exposure to polycyclic aromatic hydrocarbons. Keywords: Graphene oxide / Liquid chromatography with mass spectrometry / Magnetic solid-phase extraction / Monohydroxy polycyclic aromatic hydrocarbons / Urine DOI 10.1002/jssc.201400363
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Polycyclic aromatic hydrocarbons (PAHs) are a common carcinogen in mammals, which are mainly from incomplete combustion of organic matrices. Exposure to these environmental contaminants and their potential damage to public health has become a worldwide concern [1–3]. Once PAHs in the polluted air are absorbed into human body, they can be converted to epoxide intermediates, such as diol-epoxide, which is a highly reactive metabolite endangering cellular DNA [4]. These intermediates are the foundation of carcinogenic substances and the potential risk to human health is a chronic and accumulated process. It has been proved that exposure to PAHs may disrupt sperm DNA [5–8] and interCorrespondence: Dr. Hui Xu, Luoyu Road 152, Wuhan, Hubei Province, China E-mail: [email protected] Fax: +86-27-67867955
Abbreviations: EF, enrichment factor; MSPE, magnetic solidphase extraction; MNP, magnetic nanoparticle; PAH, polycyclic aromatic hydrocarbon C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
fere with human male fertility [9]. It has also been demonstrated that exposure to fumes and aerosols of bitumen may contribute to increased DNA strand breakage in white blood cells [10]. At present, urinary monitoring has become an increasingly relevant tool for evaluating human exposure to chemicals and pollutants [11–14], because it has many advantages, such as no invasion, pervasiveness, and accessibility. Analytical capabilities are at the core of biomonitoring for contaminants in complex urine matrices. Many analytical methods have been developed for the measurement of urinary metabolites. However, several main problems still challenge the existing analytical method of human biomonitoring. The low concentrations of metabolites in urine might be beyond instrumental LODs, and the coexisting compounds (organic solutes, inorganic ions, and macromolecules) in urine might lead to matrix interference. Therefore, an appropriate sample preconcentration step before chromatographic and mass spectrometric analysis is definitely necessary. Conventional preconcentration methods include LLE [15] and SPE [16]. Colour Online: See the article online to view Figs. 1–4 in colour. www.jss-journal.com
2592
J. Sep. Sci. 2014, 37, 2591–2598
L. Zhu and H. Xu
However, these two methods suffer from disadvantages as they are labor intensive, time consuming, and solvent consuming. SPME, a solvent-free extraction procedure, was introduced by Pawliszyn in 1990 [17]. It is a simple, fast, and green sample preparation technique, but the fiber is fragile and expensive. A new magnetic solid-phase extraction method (MSPE) based on magnetic nanoparticles (MNPs) was developed and had received increasing interest in recent years [18,19]. MNPs have high extraction efficiency because of the high surface area-to-volume ratio compared with other adsorbents. They exhibit the advantage of rapid separation from matrix due to the excellent magnetic response. MSPE has been widely applied to analysis of drugs [20, 21], environmental water samples [22,23], and biological samples [24,25]. Besides the superior magnetic response, specific surface modification, high binding capacity, and extraction efficiency are also vital for magnetic adsorbents. Therefore, many studies are focused on the surface modifications of MNPs [26–29]. For example, alkyl silane modified MNPs (C8 -amine-functionalized MNPs) were prepared and used to determine aldehyde in human urine [26]. Dodecylbenzene sulfonate coated magnetite nanoparticles were used as an adsorbent for the determination of trace amounts of ammonium in water sample [27]. Water-dispersible, photocatalytic Fe3 O4 @TiO2 coreshell magnetic nanoparticles were prepared for capturing and photocatalytically destroying some endocrine-disrupting chemicals [28]. A recent review about the application of functionalized magnetic nanoparticles in sample preparation has been reported by Xie [29]. Graphene (G) and graphene oxide (GO) are emerging and promising carbon-based nanomaterials, they have sparked great interest of scientists due to their unique 2D planar monolayer structure, superior mechanical strength, remarkable thermal and chemical stability, and ultrahigh specific surface area. For GO, the abundant oxygen-containing hydrophilic functional groups on the surface and edges can provide specific adsorption sites and stable dispersion in water. These properties make GO very attractive for various applications including sample preparation and a relevant review has been reported recently [30]. However, the direct use of GO in powder form may be inconvenient in handling and cleanup, leakage and blocking may occur when nanosized GO is used as a sorbent in SPE. Magnetic graphene-based MSPE can solve these problems effectively. Moreover, the recovery of nanoparticles from suspension can be realized easily by a magnet. Novel materials have been prepared and applied for the analysis of neonicotinoid pesticides from pear and tomato samples [31], phthalate esters from water and beverages [32] and soybean milk samples [33]. In this paper, magnetic graphene oxide composite was prepared and applied for the determination of PAH metabolites in human urine. Three polycyclic aromatic hydrocarbons, 2-naphthol, 3-hydroxy-phenanthrene, 1-hydroxypyrene were selected as model compounds. The feasibility of Fe3 O4 /GO nanocomposite as MSPE adsorbent in biomonitoring was estimated. Based on the MNPs, an MSPE–LC–MS C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
method was established. The influence of sorbent amount, extraction conditions, and desorption conditions were optimized. The developed method was applied to the analysis of human urinary OH-PAHs in two different cities of Hubei Province, China.
2 Materials and methods 2.1 Apparatus An Alliance e2695 separations module, as the HPLC system (Waters, Manchester, UK) and Massllynx 4.1 software were used to achieve the separation of the three PAH metabolites. HPLC analysis was obtained by a Waters symmetry C8 column [150 mm × 2.1 mm (id), 3.5 m] at the flow rate of 0.2 mL/min and the column temperature was 30⬚C. To get the best chromatographic separation, ultrapure water containing 5 mM/L ammonium acetate (A) and acetonitrile (B) were selected as a gradient LC system. The linear gradient elution was as follows: initial state was 50% of A, and then linearly increased to 100% within 4 min; and got back to 50% in 4 min, and equilibrated for 3 min. Injection volume was 10 L. The MS system was an Acquity TQD instrument (Waters, Manchester, UK), which was equipped with a Z-spray source for negative ESI. Selective ion reaction was executed through MS mode. Some other MS parameters were set as follows, capillary voltage: 2.5 kV, extraction voltage: 2 V, ion source temperature: 120⬚C, solvent temperature: 350⬚C, solvent gas flow: 550 L/h. The confirmation and quantification ions (m/z) were 143.09, 193.11, and 217.11 for 2-naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene, respectively.
2.2 Reagents 2-Naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene were obtained from Dr. Ehrenstorfer (Germany). HPLCgrade ethanol was purchased from Tedia (USA). HPLCgrade methanol and acetonitrile were obtained from Merck (Darmstadt, Germany). Ultrapure water was used in all experiments (Millipore Simplicity 185, USA). Glucuronidase/arylsulfatase was obtained from Merck (Darmstadt, Germany). Iron(III) chloride hexahydrate (FeCl3 ·6H2 O), iron chloride tetrahydrate (FeCl2 ·4H2 O), nitric acid (HNO3 ), hydrogen peroxide (H2 O2 ), and potassium permanganate (KMnO4 ) were all of analytical grade and purchased from Shanghai chemical reagents company (Shanghai, China). Graphite powder (325 mesh, 99.95%) was purchased from Jinrilai graphite (Qingdao, China). Other chemical reagents were all of analytical grade.
2.3 Preparation of adsorbent material Fe3 O4 nanoparticles were generated with a modified Massart method (for details see Supporting Information S2) [34]. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2591–2598
Expanded graphite was used to create graphite oxide according to a modified Hummer’s method (for details see Supporting Information S2) [35]. The synthesis of Fe3 O4 /GO nanocomposite is illustrated in Fig. 1. First, the prepared Fe3 O4 nanoparticles were dispersed in 1 mol/L HNO3 to obtain a positive surface charge. Second, graphite oxide was exfoliated to graphene oxide by ultrasound for an hour. Last, the modified Fe3 O4 and GO were mixed and mechanically agitated for 4 h to generate Fe3 O4 /GO composites [36].
Sample Preparation
2593
tration of 5 g/L. After vortex dispersion, 500 L sorbent was added in the 2 mL standard solution. Second, the mixed suspension was agitated thoroughly on a vortex mixer for 3 min to reach adsorption equilibrium. Immediately, the Fe3 O4 /GO was isolated from the sample solution with a magnet. Third, 50 L ethanol and 50 L acetonitrile were used to elute the target analytes from the sorbent (3 min). Finally, an aliquot of 10 L filtered supernatant was injected for LC–MS analysis.
3 Results and discussion 2.4 Characterization of synthetic materials
3.1 Characterization of Fe3 O4 /GO nanocomposites
Powder X-ray diffraction patterns were obtained on a Bruker D8 Advance X-ray diffractometer with CuK␣ radiation (k = 1.54178 Å). SEM images were obtained on a LEO 1450VP scanning electron microscope. Zeta potential measurements were performed using a nanoparticle zeta potential analyzer (Zen3690, Malvern, UK), and the graphene oxide and Fe3 O4 samples were diluted to 1 mg/mL before measurements. The specific surface area was obtained on high-precision BELSORP-mini specific surface area and aperture determinator (BEL JAPAN). Magnetic properties of the materials were characterized by a PPMS-9 vibrating sample magnetometer (QUANTOM, USA).
The morphology and structure of the prepared Fe3 O4 nanoparticles, graphene oxide, and Fe3 O4 /GO nanocomposites were examined by SEM and TEM (Fig. 2A–C). From Fig. 2A and A1, we can see that Fe3 O4 nanoparticles with the average diameter of about 20–30 nm were aggregated together. For GO, a planar monolayer structure with the smooth, wrinkle-like and silk-like morphology was observed (Fig. 2B and B1). SEM and TEM images of the hybrid composite confirmed the attachment of Fe3 O4 nanoparticles onto the GO surface (Fig. 2C and C1). XRD was used to characterize the phase structure of Fe3 O4 , graphite oxide, and Fe3 O4 /GO (Fig. 2D–F). It was found that the Fe3 O4 phase (JCPDS file No. 26–1136) was synthesized. The position of the diffraction peaks at 2 = 36.74, 58.16, and 64.19⬚ can be assigned to (311), (511), and (440) reflections, respectively (Fig. 2D). The graphite oxide (Fig. 2E) presented a very sharp diffraction peak at 2 = 9.29⬚, which indicated that the graphite oxide was successfully synthesized. The diffraction peak at 2 = 9.29⬚ can be assigned to the (002) reflection. In Fig. 2F, the characteristic peaks of Fe3 O4 remained unchanged, but the diffraction peak for graphene oxide increased to 2 = 30.11⬚, which indicates the conversion from graphite oxide to graphene oxide after ultrasonic treatment, a similar change has been observed previously [36]. In addition, the surface electric charges of Fe3 O4 and the GO were also measured, the result showed that the zeta potentials of acidulated Fe3 O4 and GO were 11.9 and −29.6 mV, respectively. Fe3 O4 nanoparticles dispersed in 1 mol/L HNO3 possessed positive surface charge. The surface of the as prepared GO sheets were highly negatively charged, which was apparently a result of ionization of the carboxylic acid and phenolic hydroxyl groups on the surface [36]. The data suggest that electrostatic interaction might play an important role in their combination. These data presented above indicated that the Fe3 O4 /GO composites were successfully prepared based on a simple one-step surface charge-driven self-assembly. The magnetic properties of Fe3 O4 and Fe3 O4 /GO were also determined by vibrating sample magnetometry. As shown in Supporting Information Fig. S2, we can see that the composites were superparamagnetic. The specific saturation magnetization of Fe3 O4 , Fe3 O4 /GO (10:1 m/m) and Fe3 O4 /GO (5:1 m/m) was 65.98, 58.78, and 40.53 emu/g, respectively. The surface area of synthetic Fe3 O4 /GO nanocomposites tested by nitrogen
2.5 Preparation of standard solution Two types of standard solution were prepared. Stock solutions of 1000 mg/L 2-naphthol, 500 mg/L 3-hydroxyphenanthrane, and 1000 mg/L 1-hydroxypyrene were prepared separately in methanol, stored at −20⬚C. The daily standard working solution was prepared by dilution and mixing the stock solution according to demand. The calibration solution used in method validation was prepared by spiking the standard solution into real urine matrix (the signals of real urine were subtracted from those of the spiked urine).
2.6 Pretreatment of urine samples Thirty eight urine samples were collected in the morning from nonsmoking men aged 18–22. Before use, urine samples were filtered with a 0.22 m filter membrane. Ten microliters of -glucuronidase/arylsulfatase and a certain amount of standard solution were added to 10 mL of the urine pool (pH 5.5) in a 12 mL glass vial. Then, the mixture was incubated at 37⬚C for 4 h away from light.
2.7 Magnetic solid-phase extraction procedure The MSPE process is as follows (Supporting Information Fig. S1). First, Fe3 O4 /GO nanocomposites were dispersed into deionized water to create a suspension with the concen C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
2594
L. Zhu and H. Xu
J. Sep. Sci. 2014, 37, 2591–2598
Figure 1. Schematic illustration of the fabrication of Fe3 O4 /GO nanocomposite.
adsorption by the BET method was 109.58 m2 /g. The large surface area of composites provided adequate adsorption sites for the following extraction application.
3.2 Selection of extraction conditions The mass ratio between Fe3 O4 and GO was a crucial factor in the extraction. Because iron content has a great impact on the magnetism of synthetic materials. A higher Fe3 O4 percent-
age means more rapid magnetic separation. As shown in the Supporting Information Fig. S2, Fe3 O4 /GO (10:1 m/m) has a stronger superparamagnetism than Fe3 O4 /GO (5:1 m/m), the results indicated that increased loading amounts of Fe3 O4 can ensure the stronger superparamagnetism of nanocomposites. While for GO, a higher ratio means a larger surface area, more adsorption sites, and better dispersion in aqueous solution. In order to investigate the adsorptive capabilities of Fe3 O4 /GO composites with different proportions for PAHs, five kinds of nanocomposites (1:1, 5:1, 10:1, 15:1, 20:1,
Figure 2. SEM images (A, B, C) and XRD patterns (D, E, F) of the freshly prepared nanoparticles. Insets are TEM images (A-1, B-1, C-1). (A) and (D): Fe3 O4 , (B): GO, (C) and (F): Fe3 O4 /GO, (E): Graphite oxide. Scale bars for A, B, C, A-1, B-1, and C-1 are 10 nm, 1 m, 100 nm, 50 nm, 1 m, and 100 nm, respectively.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2591–2598
Sample Preparation
2595
Figure 3. Effect of the extraction time (A) and desorption time (B) on the peak areas. Conditions: 2 mL sample solution, concentration of 2-naphthol, 3-hydroxyphenanthrene, and 1-hydroxypyrene: 25, 12.5, and 25 ng/mL, Fe3 O4 /GO absorbent: 500 L, 2 g/L.
500 L, 2 g/L) were added into 2 mL aqueous solution containing 25 ng/mL of 2-naphthol, 12.5 ng/mL of 3-hydroxyphenanthrene and 25 ng/mL of 1-hydroxypyrene. The best result was obtained at the ratio of 10:1 (Fe3 O4 /GO, m/m, for details see Supporting Information Fig. S3a). Meanwhile, the extraction efficiency of bare Fe3 O4 was also investigated. The result showed that bare Fe3 O4 can also adsorb OH-PAHs, while the extraction efficiency was far less than those of Fe3 O4 /GO. Thus, it can be concluded that the adsorption of OH-PAHs is ascribed mainly to GO. In addition, the effect of the Fe3 O4 /GO amount on the extraction efficiency was further investigated. Nanocomposites at the concentration of 5 g/L showed the best extraction efficiency (Supporting Information Fig. S3b). Higher concentration means more adsorbents, which is beneficial for extraction, but it is inconvenient for the withdrawing of desorption solvent in desorption procedure. Accordingly, 5 g/L of 10:1 Fe3 O4 /GO was used as adsorbent in the following experiments. The mixture model of absorption and desorption was also investigated. The result showed that vortex was the best way for this work, because vortex can accelerate the mass transfer process of analytes from sample solution to dispersive material. Furthermore, the effect of adsorption time was studied in the range of 1 to 6 min. As the extraction time profile shows in Fig. 3A, there was no significant difference in the peak area of 2-naphthol and 1-hydroxypyrene with time. For 3-hydroxy-phenanthrene, the best extraction efficiency was obtained at 4 min. The results indicated that 4 min was adequate to perform the extraction. The rapid extraction time was attributed to high surface area provided by the nanosized magnetic sorbents. Thus, we chose 4 min as the optimum extraction time.
3.3 Selection of desorption conditions An effective elution can ensure high recovery of extraction method. In our work, five different organic solvents C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(methanol, ethanol, acetonitrile, methanol/acetonitrile, and ethanol/acetonitrile) were selected as eluents and their desorption efficiencies were investigated. The results showed that the mixture of acetonitrile and ethanol had a higher desorption capacity than other solvents (Supporting Information Fig. S4a). And, the ratio of acetonitrile to ethanol (2:8, 3:7, 5:5, 7:3, 8:2, v/v) was studied further, the results showed that the mixed solvent at the ratio of 1:1 offered better desorption efficiency (Supporting Information Fig. S4b). The desorption time profile was also investigated, as can be seen from Fig. 3B, the maximum peak signal was obtained at 3 min for 2-naphthol and 3-hydroxyphenanthrene. While for 1-hydroxypyrene, the signal decreased continuously with time. Hence, 3 min was selected as the appropriate desorption time. The eluent volume was also optimized, 100 L was selected to ensure high enrichment and recovery. In summary, analytes were vortex extracted for 4 min with 5 g/L of Fe3 O4 /GO (10:1, m/m) nanocomposites as adsorbent and desorbed in 100 L acetonitrile and ethanol mixture (1:1, v/v) for 3 min in the following investigation. Under optimal conditions, the enrichment factors of three target compounds were calculated. Enrichment factor (EF) is defined as the ratio of the peak area of an extracted analyte to that in the original sample determined by direct injection of the solution. The data for 2-naphthol, 3hydroxyphenanthrene, and 1-hydroxypyrene were 5.4, 17.1, and 19.5, respectively (theoretical enrichment factor: 20). LogP (octanol/water partition coefficient, an indicator for hydrophobicity) values were 2.71, 3.909, and 4.43 for three analytes. The Fe3 O4 /GO composites exhibited high EF values for 3-hydroxy-phenanthrene (17.1) and 1-hydroxypyrene (19.5), suggesting the high extraction selectivity toward hydrophobic compounds, the similar results were also reported in previous report [37]. It is well known that there exists the – stacking interaction between GO and the benzene ring of PAHs, and the interaction increases with the ring number of PAHs [38–40]. It was reported that the graphene oxide is more likely to adsorb the targets with more benzene rings. This order of EFs was also consistent with the ring number www.jss-journal.com
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
103.6 125.2 99.7 12.6 9.8 11.1 6.8 9.9 9.8 15.5 9.3 8.2 1–3000 0.2–300 0.1–300 y = 208.16x + 4715.9 y = 5072.4x + 28034 y = 10461x + 83796 2-Naphtholb) 3-Hydroxy-phenanthrene 1-Hydroxypyrene
a) The concentrations unit of analytes in RSD and recovery experiments was ng/mL. b) The low, medium, high concentrations of 2-naphthol were 1, 10, and 100 ng/mL in RSD experiments, respectively.
13.4 14.0 10.2
10 1
12.0 9.6 6.8 7.1 13.1 8.5 0.15 0.13 0.01
0.1 0.1
1
Inter-day RSDa) (%, n=5) Intra-day RSDa) (%, n = 5) Linear range Regression equation
To further demonstrate the advantages of the proposed method, the proposed method was compared with previous published reports, which used three conventional sample preparation methods (LLE [15], SPE [16], SPME [17]). It is clearly seen from Table 2 that our method showed obvious advantages over the other three methods in analysis time and recovery. With nanocomposites as sorbent, fast extraction can be obtained due to the high surface area-to-volume ratio of nanocomposites. High recovery data (98.3–125.2%) indicated that the satisfactory accuracy of our method. Separation of magnetic sorbents from large-volume samples is convenient and fast with the aid of magnet and no filtration and centrifugation is needed. Sensitivity of our method (LC– MS) was a little lower than that of a GC–MS method, the possible explanation is the ionization of OH-PAHs in electrospray ion sources (LC–MS) is poorer than in electron ionization source (GC–MS) due to the weak polarity of analytes. In summary, this method possesses many advantages, including rapid analysis, easy operation, satisfactory accuracy, and adequate sensitivity.
Analyte (ng/mL)
3.5 Method comparison
Table 1. Method validation data for the established MSPE-LC-MS method
r
LOD
A series of quantitative experiments in regard to linear range, correlation coefficient, and LOD, reproducibility, and recovery were conducted to examine the proposed method in urine matrix under the optimal experimental conditions. Linear regression analysis was presented using peak area against concentration of analytes. Table 1 lists the quantitative parameters of the proposed MSPE method. The LOD, which was calculated using S/N = 3, was 0.01–0.15 ng/mL. Calibration curves of three analytes exhibited good linearity with a correlation coefficient above 0.9945. The precision was investigated with five parallel experiment results of within-day precision and between-day at three concentration levels (0.1, 1, 10 ng/mL). The results in Table 1 show that RSD values ranged from 6.8 to 14.0%. Recovery experiments of MSPE procedure were performed by measuring the known concentrations analytes in urine samples of healthy people at two spiked concentrations (10, 100 ng/mL). The average recoveries were in range of 98.3–125.2% (Table 1). Satisfactory recovery demonstrated that the real urine matrix has little interference on the performance of MSPE and it could be concluded that the MSPE–LC–MS analytical method is reliable [41, 42].
10
3.4 Method validation
0.9959 0.9959 0.9945
100
Recoverya) (%)
of OH-PAHs. Based on the result, we can deduce that both – stacking interaction and hydrophobic interaction play important roles in the extraction of OH-PAHs.
115.4 98.3 117.3
J. Sep. Sci. 2014, 37, 2591–2598
L. Zhu and H. Xu
10
2596
www.jss-journal.com
Sample Preparation
J. Sep. Sci. 2014, 37, 2591–2598
2597
Table 2. Method comparison
Method
Extraction timec) (min)
LOD (ng/mL)
Recovery (%)
RSD (%)
Organic solvent
References
LLE(GC–IDHRMS)a)
>80 min
0.007–0.012
46–72
2.9–11
[15]
SPE(LC–EI-MS)
—
10–100
—
3.7–8.8
SPME(ID-GC–MS) MSPE(HPLC–MS)
>165 min 8 min
0.00078–0.0158 0.01–0.15
6–47 98.3–125
2.4–18.7 6.8–15.5
10 mL pentane; 5 L dodecane; 20 L toluene; 10 L MSTFAb) 3 mL methanol 6 mL fomic acid 15 L MSTFA 100 L ethanol and acetonitrile
[16] [17] This work.
a) Isotope dilution high-resolution MS. b) N-methyl-N-(trimethylsilyl)trifluoroacetamide. c) The time of treatment of urine is not included.
3.6 Application in real samples This proposed method was applied for the determination of three OH-PAHs in the urine of nonsmoking men aged 18–20 (n = 38) who lived in two cities with different environments (Wuhan and Enshi). The preliminary research aimed to evaluate the possibility of the method for biological monitoring of environment exposure to PAHs. With a current population exceeding nine million, the city of Wuhan is an emerging mega-city in Central China. It is also one of the most important industrial cities. And it has been reported that the air in Wuhan is highly polluted [43]. PAHs, from emission of coal-burning power plants, steel plants, or automobile exhaust are the main atmospheric pollutants in Wuhan. Enshi, a famous ecological tourist city in Central China, is surrounded by forest-covered mountains and rivers. Agriculture and tourism are the main economical drivers of the city. Wuhan and Enshi are ideal sites for the comparative study of environment pollution episodes, representative of Chinese cities with different populations and industrialization. Figure 4 depicts typical LC–MS–SIR chromatograms of blank and spiked samples of healthy volunteer. The results showed that the concentration of 1-hydroxypyrene (widely used as a kind of suitable PAH biomarker [44]) was higher than the other two analytes. Furthermore, there was a clear regional difference about the urinary concentration of 1hydroxypyrene in two cities. The average concentration of 1hydroxypyrene in place A (Wuhan, Hubei, China) was about 0.21 ng/mL and the detection rate was 84%. Contrastively, that concentration in place B (Enshi, Hubei, China) was much lower (0.12 ng/mL, the detection rate was only 68%), and some analytes cannot be detected using the proposed method. The difference in urinary concentration may be ascribed to the different exposure environment of the two cities. Compared with people in tourist city, distinctly higher urinary OHPAHs level was found in individuals who lived in industrial city. The results indicated that the urinary level of OH-PAHs could reflect the pollution level of people’s residence and the
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4. LC–MS-SIR Chromatograms of OH-PAHs in blank urine sample (A); standard solution (B, 1 ng/mL); spiked urine (C, 10 ng/mL; D, 100 ng/mL); the insert is the urinary concentration of 1-hydroxypyrene in two cities (A: Wuhan, B: Enshi).
proposed method has the potential to determine pollutants and their metabolites in human urine.
4 Conclusion A facile determination of PAH metabolites in human urine based on the magnetic graphene oxide was reported in this paper. The hybrid Fe3 O4 /GO nanocomposites exhibit excellent magnetic response and satisfactory extraction performance toward three OH-PAHs. For assay of target analytes, the proposed method possesses numerous advantages, including cost effectiveness, time saving, ease of operation, and procedural simplicity. In the urine matrix, reasonable recovery, acceptable reproducibility, and low LOD were achieved for the method. In summary, this MSPE–HPLC–MS method has potential for the biomonitoring of OH-PAHs in human
www.jss-journal.com
2598
L. Zhu and H. Xu
J. Sep. Sci. 2014, 37, 2591–2598
urine, and it can provide useful information for environmental exposure to PAHs.
[17] Smith, C. J., Walcott, C. J., Huang, W., Maggio, V., Grainger, J., Patterson, D. G. Jr, J. Chromatogr. B 2002, 778, 157–164.
This work was supported by the National Science Foundation of China (Nos. 21175052 and 21105035), and the selfdetermined research funds of the CCNU from the colleges’ basic research and operation of MOE (No. CCNU14A05010).
[18] Robinson, P. J., Dunnill, P., Lilly, M. D., Biotechnol. Bioeng. 1973, 15, 603–606. ˇ ˇ ´ M., Safaˇ [19] Safaˇ rı́kova, rı́k, I., J. Magn. Magn. Mater. 1999, 194, 108–112.
The authors have declared no conflict of interest.
[20] Wang, Y., Wang, Y., Chen, L., Wan, Q.-H., J. Magn. Magn. Mater. 2012, 324, 410–417. [21] Gao, Q., Lin, C.-Y., Luo, D., Suo, L.-L., Chen, J.-L., Feng, Y.-Q., J. Sep. Sci. 2011, 34, 3083–3091.
5 References
[22] Liu, X., Lu, X., Huang, Y., Liu, C., Zhao, S., Talanta 2014, 119, 341–347.
[1] Grimmer, G., Jacob, J., Dettbarn, G., Naujack, K. W., Int. Arch. Occ. Env. Hea. 1997, 69, 231–239.
[23] Su, S., Chen, B., He, M., Hu, B., Xiao, Z., Talanta 2014, 119, 458–466.
[2] Masuck, I., Hutzler, C., Luch, A., J. Sep. Sci. 2011, 34, 2686–2696.
[24] Zheng, H. B., Mo, J. Z., Zhang, Y., Gao, Q., Ding, J., Yu, Q. W., Feng, Y. Q., J Chromatogr. A. 2014, 1329, 17–23.
[3] Marais, A. A. S., Laurens, J. B., J. Sep. Sci. 2005, 28, 2526–2533.
[25] Moazzen, M., Ahmadkhaniha, R., Gorji, M. E. h., Yunesian, M., Rastkari, N., Talanta 2013, 115, 957–965.
[4] Shimada, T., Fujii-Kuriyama, Y., Cancer Sci. 2004, 95, 1–6.
[26] Song, D., Gu, Y., Liang, L., Ai, Z., Zhang, L., Xu, H., Anal. Methods 2011, 3, 1418–1423.
´ [5] Sanchez-Guerra, M., Pelallo-Mart´ınez, N., D´ıaz-Barriga, ´ F., Rothenberg, S. J., Hernandez-Cadena, L., Faugeron, ´ S., Oropeza-Hernandez, L. F., Guaderrama-D´ıaz, M., Quintanilla-Vega, B., Mutat. Res.-Gen. Tox. En. 2012, 742, 66–71. [6] Gurbani, D., Bharti, S. K., Kumar, A., Pandey, A. K., Ana, G. R. E. E., Verma, A., Khan, A. H., Patel, D. K., Mudiam, M. K. R., Jain, S. K., Roy, R., Dhawan, A., Int. J. Hyg. Envir. Heal. 2013, 216, 553–565. ´ [7] Cebulska-Wasilewska, A., Pawłyk, I., Panek, A., Wiechec, A., Kalina, I., Popov, T., Georgieva, T., Farmer, P. B., Mutat. Res.-Fund. Mol. M. 2007, 620, 145–154. [8] Singh, R., Sram, R. J., Binkova, B., Kalina, I., Popov, T. A., Georgieva, T., Garte, S., Taioli, E., Farmer, P. B., Mutat. Res.-Fund. Mol. M. 2007, 620, 83–92. [9] Han, X., Zhou, N. Y., Cui, Z. H., Ma, M. F., Li, L. B., Cai, M., Li, Y. F., Lin, H., Li, Y., Ao, L., Liu, J. Y., Cao, J., Environ. Health Persp. 2011, 119, 652–657. [10] Marczynski, B., Raulf-Heimsoth, M., Preuss, R., Kappler, M., Kafferlein, H. U., Schott, K., Pesch, B., Mensing, T., Spickenheuer, A., Angerer, J., Bruning, T., Toxicol. Lett. 2006, 164, S282–S282. [11] Yusa, V., Ye, X., Calafat, A. M., TrAC, Trends Anal. Chem. 2012, 38, 129–142. [12] Guo, Y., Senthilkumar, K., Alomirah, H., Moon, H. B., Minh, T. B., Mohd, M. A., Nakata, H., Kannan, K., Environ. Sci. Technol. 2013, 47, 2932–2938. [13] Feldt, T., Fobil, J. N., Wittsiepe, J., Wilhelm, M., Till, H., ¨ Zoufaly, A., Burchard, G., Goen, T., Sci. Total Environ. 2014, 466–467, 369–376. [14] del Carmen Hurtado-Sanchez, ´ M., Espinosa-Mansilla, ´ ´ A., Rodr´ıguez-Caceres, M. I., Mart´ın-Tornero, E., Duran´ I., J. Sep. Sci. 2012, 35, 2575–2584. Meras, [15] Li, Z., Romanoff, L. C., Trinidad, D. A., Hussain, N., Jones, ¨ R. S., Porter, E. N., Patterson, D. G., Sjodin, A., Anal. Chem. 2006, 78, 5744–5751. [16] Andreoli, R., Manini, P., Bergamaschi, E., Mutti, A., Franchini, I., Niessen, W. M. A., J. Chromatogr. A. 1999, 847, 9–17. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[27] Eskandari, H., Shariati, M. R., Anal. Chim. Acta 2011, 704, 146–153. [28] Chalasani, R., Vasudevan, S., ACS Nano. 2013, 7, 4093–4104. [29] Xie, L., Jiang, R., Zhu, F., Liu, H., Ouyang, G., Anal. Bioanal. Chem. 2014, 406, 377–399. [30] Liu, Q., Shi, J., Jiang, G., TrAC, Trends Anal. Chem. 2012, 37, 1–11. [31] Ma, X., Wang, J., Sun, M., Wang, W., Wu, Q., Wang, C., Wang, Z., Anal. Methods 2013, 5, 2809–2815. [32] Wu, Q., Liu, M., Ma, X., Wang, W., Wang, C., Zang, X., Wang, Z., Microchim. Acta. 2012, 177, 23–30. [33] Wang, W., Ma, R., Wu, Q., Wang, C., Wang, Z., Talanta 2013, 109, 133–140. [34] Massart, R., IEEE T. Magn. 1981, 17, 1247–1248. [35] Hummers, W. S., Offeman, R. E., J. Am. Chem. Soc. 1958, 80, 1339–1339. [36] Han, Q., Wang, Z., Xia, J., Chen, S., Zhang, X., Ding, M., Talanta 2012, 101, 388–395. [37] Fan, J., Dong, Z., Qi, M., Fu, R., Qu, L., J Chromatogr. A 2013, 1320, 27–32. [38] Zhao, G., Jiang, L., He, Y., Li, J., Dong, H., Wang, X., Hu, W., Adv. Mater. 2011, 23, 3959–3963. [39] Zhao, G., Li, J., Wang, X., Chem. Eng. J. 2011, 173, 185– 190. [40] Wu, T., Cai, X., Tan, S., Li, H., Liu, J., Yang, W., Chem. Eng. J. 2011, 173, 144–149. [41] Yan, S., Qi, T.-T., Chen, D.-W., Li, Z., Li, X.-J., Pan, S.-Y., J Chromatogr. A 2014, 1347, 30–38. [42] Farajzadeh, M. A., Mogaddam, M. R. A., Ghorbanpour, H., J Chromatogr. A 2014, 1347, 8–16. [43] Qian, Z., He, Q., Lin, H.-M., Kong, L., Liao, D., Dan, J., Bentley, C. M., Wang, B., Environ. Res. 2007, 105, 380–389. [44] Jacob, J., Seidel, A., J. Chromatogr. B. 2002, 778, 31–47.
www.jss-journal.com
