Restricted-access nanoparticles for magnetic solid-phase extraction of steroid hormones from environmental and biological samples

Journal of Chromatography A, 1244 (2012) 46–54

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Restricted-access nanoparticles for magnetic solid-phase extraction of steroid hormones from environmental and biological samples Lei Ye a , Qing Wang a , Jinping Xu a , Zhi-guo Shi b , Li Xu a,∗ a b

Tongji School of Pharmacy, Huazhong University of Science and Technology, Wuhan 430030, China Department of Chemistry, Wuhan University, Wuhan 430072, China

a r t i c l e

i n f o

Article history: Received 10 January 2012 Received in revised form 6 April 2012 Accepted 29 April 2012 Available online 7 May 2012 Keywords: Restricted-access material Magnetic nanoparticle Magnetic solid-phase extraction Non-ionic surfactants

a b s t r a c t Restricted-access materials based on non-ionic surfactant-coated dodecyl-functionalized magnetic nanoparticles were prepared and applied to extract steroid hormones from environmental and biological samples. The magnetic nanoparticles were synthesized by co-precipitation, and were functionalized with dodecyltriethoxysilane, giving dodecyl-grafted magnetic nanoparticles (C12 -Fe3 O4 ). They were further modified with different non-ionic surfactants by self-assembly adsorption. Several types of non-ionic surfactants, Tween-20, 40, 60 and 85, and Span-40, 60 and 80, were investigated as the coatings. Tween surfactants coated C12 -Fe3 O4 , named as TW-20 (40, 60, 85)-C12 , exhibited good dispersibility in aqueous solution, which was a preferred character in extraction; besides, TW-20-C12 and TW-40-C12 showed good anti-interference ability and satisfactory reproducibility when they were used as magnetic solid-phase extraction (MSPE) sorbents. The factors that may influence the extraction, including the amount of magnetic nanoparticles, extraction and desorption time, the amount of salt addition, the type and volume of desorption solvent, the volume of methanol addition and pH of sample solution, were investigated in detail. High performance liquid chromatography–UV detection was employed for analysis of target analytes (steroid hormone compounds). The developed method was successfully used for the determination of the target analytes in environmental and urine samples. Both tested materials afforded good recovery, satisfactory reproducibility and low limits of detection for environmental samples, which indicates that the materials possessed anti-interference ability. However, compared to TW-40-C12 , TW-20-C12 nanoparticles provided better recovery in relatively complex biological samples, which may indicate that the latter one is more appreciated in complex samples. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Analysts encounter an ongoing challenge to pursue efficient analysis in complex samples, due to low concentration of analytes and unavoidable interference of complex matrices to instrumental analysis. As a result, an appropriate sample pretreatment is necessary in analysis, which aims at cleaning up, isolating and/or concentrating analytes of interest. In recent years, to afford operation convenience, enhance the extraction efficiency and save the cost, magnetic solid-phase extraction (MSPE) has gained much attention [1–3]. In MSPE, magnetic nanoparticles (MNPs), generally Fe3 O4 or ␥-Fe2 O3 , are used as sorbents. Compared to traditional SPE sorbents, MNPs possess high surface area and have unique magnetic properties. The equilibrium between the sorbents and sample solutions can be reached quickly once MNPs are introduced to the solution. After extraction, MNP sorbents enriched with analytes are easily isolated from suspension

∗ Corresponding author. Tel.: +86 27 8369 2735; fax: +86 27 8369 2762. E-mail address: [email protected] (L. Xu). 0021-9673/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2012.04.075

by exerting an exterior magnetic force. Hence, unlike traditional cartridge SPE, MSPE avoids column packing and possible blockage during use. However, bare MNPs were observed to aggregate easily, which may alter their stability and extraction capacity [4,5]. Derivatization has been a key to overcome some weakness of MNPs [6–8]. Recently, Chen et al. [9] reviewed applications of derivatized MNPs, e.g. silanized- and polymer modified-MNPs, in water samples. Nevertheless, modified MNPs still suffer from some defects. For instance, C18 -MNPs were difficult to disperse in water samples due to their high hydrophobicity and may lose their adsorption ability in complex matrices as they could be easily contaminated by sample matrix [6,8]. Hence, appropriately functionalized materials, being compatible with sample matrix, are still desired in MSPE. Restricted access material (RAM) is a class of biocompatible SPE sorbents. Generally, RAM has interior support suitable for small molecules’ retention and exterior surface with different groups to exclude macromolecules to the interior support. Therefore, it weakens or eliminates adsorption of uninterested macromolecules (e.g. proteins) without negative influence on extraction of target analytes, and a reliable cleanup of complex matrices can be achieved. Since its discovery, RAM has been satisfactorily used in

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

environmental and biological analysis [10–13]. However, to our best knowledge, there are few papers drawing attention to RAM combining with MNPs. Cai et al. synthesized chitosan-coated C18 MNPs to extract phthalate ester and perfluorinated compounds from environmental samples. The chitosan coating was demonstrated to have anti-interference ability to macromolecules by size exclusion or electrostatic repulsion [6,8]. Additionally, magnetic microspheres with mesoporous shell were reported to exclude proteins by size exclusion [14,15]. Very recently, MNPs modified with diol groups were also reported to generate RAM [16]. In previous work, ionic surfactants, e.g. cetyltrimethylammonium bromide (CTAB), octadecyltrimethylammonium (OTAB) and sodium dodecyl sulfate (SDS), were used to form hemimicelles and/or admicelles on the surface of MNPs by means of hydrophobic and/or electrostatic interactions [7,17]. The ionic surfactants provided functionality, e.g. hydrophobic and electrostatic groups, for extraction. Polyoxyethylene-containing non-ionic surfactant was reported to have anti-interference ability when used as SPE or high performance liquid chromatographic (HPLC) sorbent coating [18,19]. However, there is a lack of studies about non-ionic surfactant coating in MSPE. In the present study, RAM-MSPE based on non-ionic surfactant coating was developed. Commonly used non-ionic surfactants, e.g. Tween- and Span-series, were chosen as the MNP coatings, which were supposed to shield macromolecules in complex sample matrix. Prior to coating, the MNPs were derivatized with C12 , which may interact with the target analytes. Steroid hormone compounds (ST-HRMs), a group of biological chemicals derivatized from cholesterol, were chosen as test analytes due to their potential negative effect on ecospecies and human beings. 2. Experimental 2.1. Chemicals and reagents Four ST-HRMs, hydrocortisone (HC), 4-androstene-3,17-dione (AD), progesterone (PG) and testosterone propionate (TP) were obtained from Aladdin Chemical Reagent Co., Ltd. (Shanghai, China). Tween-20, 40, 60 and 85, and Span-40, 60 and 80, HPLC-grade methanol (MeOH), acetonitrile (ACN) and NaCl were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Acetone was purchased from Concord Co., Ltd. (Tianjin, China). HCl was from Kaifeng Dong Da Chemical Reagent Co., Ltd. (Henan, China). Dodecyltriethoxysilane was obtained from Hubei Wuhan University Silicone New Material Co., Ltd. (Wuhan, China). FeCl2 ·4H2 O was purchased from Shanghai Gongxuetuan No. 2 Experiment Factory (Shanghai, China). Ultrapure water (pH 6.2) was produced by a Heal Fore NW system (Shanghai, China). 2.2. Apparatus The determination of the four ST-HRMs was performed on a Hitachi (Tokyo, Japan) HPLC system. It consists of a Model L2130 pump, a Rheodyne 7725i injector (Cotati, CA, USA) and an L-2400 UV–vis spectrophotometric detector. Data were collected and processed by T3000P (Hangzhou Hui Pu Technology Co., Ltd., Hangzhou, China) software. Chromatographic separation was performed on an ODS (4.6 × 250 mm, 5 ␮m) column from Shimadzu (Kyoto, Japan) at a temperature of 22 ◦ C. A mixture of MeOH and water (78:22, v:v) was used as mobile phase. The flow rate was 1.2 mL min−1 and the injection volume was 20 ␮L. The UV wavelength was set at 242 nm. All the experiments were performed at least in triplicate. The pH values were measured with a Mettler Toledo Delta 320 pH meter (Shanghai, China). Fourier transform infrared

47

spectroscopy (FT-IR) was performed on AVATAR 360 (Thermo, USA). Thermogravimetric analysis (TGA) was carried out using a Setaram (France) TG-DTA analyzer. Transmission electron microscope (TEM) images were obtained on FEI Tecnal G2 12 (Netherlands). 2.3. Sample preparation Stock solutions of the four ST-HRMs (0.5 mg mL−1 of each analyte) were prepared separately in MeOH. They were stored at 4 ◦ C. Water samples were prepared by spiking ultrapure water with the analytes at a known concentration (0.5 ␮g mL−1 ) to study extraction performance under different conditions. The genuine water samples were collected from Hanjiang River, Changjiang River and East Lake (Wuhan, Hubei, China). Tap water samples were from laboratory (Tongji School of Pharmacy, Huazhong University of Science and Technology, Wuhan, China). They were stored in brown glass containers at the temperature of 4 ◦ C. Prior to extraction, they were adjusted to a desired pH without any other pretreatment. The urine samples were collected from volunteers (both male and female). Before extraction, the urine was spiked with the stock solutions to give a desired concentration and diluted with pure water (1:10, v:v). The above solution was then adjusted to the desired pH and was subjected to the extraction process without any other pretreatment. For the blank urine sample, it was subjected to the same sample pretreatment procedures as the spiked one. All the samples were freshly prepared daily. Each sample (4 mL) was used for extraction. The relative recoveries were determined as ratios of HPLC signal of the spiked extracted samples (real matrices) to that of standard solution at the same comparable concentrations. 2.4. Preparation of surfactant-coated MNPs The preparation scheme of surfactant-coated C12 -MNPs is shown in Fig. 1. (1) Fe3 O4 MNPs were prepared by an oxidativecoprecipitation method [20] and dried at 80 ◦ C for 24 h. (2) Fe3 O4 MNPs were functionalized with dodecyltriethoxysilane. Briefly, Fe3 O4 MNPs (2.0 g) were dispersed in the mixture of anhydrous toluene (50 mL) and dodecyltriethoxysilane (10 mL) by ultrasonication for 20 min. Then, the above mixtures reacted at 120 ◦ C for 16 h. The products, named as C12 -MNPs, were separated from the resulting solution by a magnet, washed with copious MeOH and dried for the following use. (3) Non-ionic surfactant modification was carried out at room temperature by self-assembly adsorption. For Tweenseries surfactants (Tween-20, 40, 60, 85), 2.0 g C12 -MNPs were dispersed in 150 mL water containing desired surfactant (0.5%, w/w) under vigorous mechanical stirring for 4 h. Tween-20, 40, 60, 85 coated C12 -MNPs were named as TW-20-C12 , TW-40-C12 , TW-60-C12 and TW-85-C12 , respectively. For Span-series surfactants (Span-40, 60, 80), 2.0 g C12 -MNPs were dispersed in 150 mL water–MeOH (7:1, v:v) containing desired surfactant (0.5%, w/w) under vigorous mechanical stirring for 4 h. Span-40, 60, 80 coated C12 -MNPs were named as SP-40-C12 , SP-60-C12 and SP-80-C12 , respectively. Finally, the above MNPs were washed with pure water and dried at 80 ◦ C for 12 h. 2.5. MSPE procedure The procedure of MSPE was as follows. Certain amount of NaCl was dissolved in the sample solution (4 mL), which was contained in a 5 mL vial. Prescribed amount of surfactant-C12 -MNPs was then added into the above solution. After being sonicated for certain time, the MNPs were isolated from the suspension with a Nd–Fe–B magnet (50 mm × 50 mm × 10 mm), followed by adding ultrapure water (500 ␮L) into the vial to wash the remnant sample solution.

48

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

Fig. 1. Schematic presentation of a synthetic route of surfactant-coated C12 -MNPs.

Subsequently, certain volume of eluant solvent was introduced to desorb the adsorbed analytes from the MNPs by sonication for prescribed time. The eluate was separated from the suspension with the magnet. Finally, 20 ␮L of the eluate was injected into the HPLC system for analysis. 3. Results and discussion 3.1. Selection of surfactant coating A suitable coating is essential to overcome weakness of MNPs and affords functionalities for MSPE. In this study, non-ionic surfactants were chosen for several reasons. Firstly and most importantly, non-ionic surfactants, e.g. Tween-series, have been demonstrated to form a porous and reticular film on the HPLC sorbents as RAM coatings [18]. When they were used to coat C12 -MNPs, a RAM was highly expected. It would retain low-mass analytes by hydrophobicity, but limit the access of macromolecules by the exterior surfactant coating. Secondly, water-soluble non-ionic surfactant

coating on C12 -MNPs could improve the dispersibility of MNPs in aqueous solution, which is beneficial for extraction. Additionally, compared with mixed-hemimicelle MNPs [7,21–23], non-ionic surfactants-coated MNPs were supposed to be stable in wider pH range since hydrophobic interaction rather than electrostatic interaction was predominantly involved. The extra merit may be that operation is convenient, because, unlike mixed-hemimicelle MNPs, no extra reagent was required to strengthen the interactions between the surfactants and the support. In our preliminary study, non-ionic surfactants, including water-soluble series (e.g. Tween-20, 40, 60 and 85) and fatsoluble series (e.g. Span-40, 60 and 80), were investigated to coat C12 -MNPs, respectively. The anti-interference ability of these materials was evaluated using three genuine environmental samples (Changjiang River, Hanjiang River and East Lake) and one biological sample (human urine) in terms of recovery, as compared in Table 1. Generally, TW-C12 -MNPs showed better performance than SP-C12 -MNPs and C12 -MNPs. SP-C12 -MNPs and C12 -MNPs were difficult to disperse in aqueous samples because obvious interface

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

49

Table 1 Relative recoveries of ST-HRMs in spiked real samples extracted with different non-ionic surfactant coated MNPs. Sorbents

Relative recoverya (%) Changjiang River

C12 -Fe3 O4 TW-20-C12 TW-40-C12 TW-60-C12 TW-85-C12 SP-40-C12 SP-60-C12 SP-80-C12

Hanjiang River

East Lake

Human Urine

HC

AD

PG

TP

HC

AD

PG

TP

HC

AD

PG

TP

HC

AD

PG

TP

75.4 73.6 111.8 69.2 133.7 68.6 81.5 –

140.8 111.0 124.2 83.4 148.6 66.1 76.9 –

248.8 104.2 116.9 113.8 134.0 96.3 93.1 –

446.7 107.0 106.7 178.7 103.7 108.4 206.4 –

82.3 125.1 123.4 52.8 239.4 27.6 53.8 –

126.5 123.8 117.7 61.3 327.9 27.8 63.7 –

175.1 134.4 121.1 81.3 192.2 71.9 75.5 –

204.0 127.8 106.1 57.6 79.0 100.0 365.6 –

61.8 119.1 69.4 145.9 150.7 27.5 53.9 –

91.8 146.3 83.5 154.4 215.3 27.8 76.8 –

156.5 133.0 88.9 158.1 181.5 68.4 108.2 –

320.2 122.6 87.8 205.6 241.2 90.9 231.7 –

ndb 73.7 174.0 nd nd 8.7 209.2 –

10.4 81.6 55.6 48.3 31.4 11.4 378.5 –

8.9 81.6 70.7 78.9 63.7 42.6 85.3 –

10.9 69.2 88.6 176.5 119.2 101.7 470.0 –

Extraction condition: Four milliliter of genuine sample spiked with 0.5 ␮g mL−1 of each target analyte was studied. The pH of the sample solution was adjusted to 7.0 in the presence of 200 mg mL−1 NaCl. Twenty milligram of sorbents were used. a The relative recoveries were determined as ratios of HPLC signal of the spiked extracted samples to that of standard solution at the same comparable concentrations. b Not detected.

was observed between materials and sample solution once they were mixed. Especially, in the case of SP-80-C12 , uneven baseline of HPLC was observed, interfering with the separation of analytes and resulting in unreliable data. This may be due to unstable SP-80 coating. In addition, TW-60-C12 and TW-85-C12 showed discontented recoveries. For example, TW-60-C12 provided recoveries in the range of 146–206 in Hanjiang River; while, TW-20-C12 and TW-40-C12 afforded more reasonable recoveries than the others, and were thus supposed to have good anti-interference abilities. These observations were related to different properties of surfactants. Hydrophile–lipophile balance (HLB) values are 16.7 and 15.6 for Tween-20 and Tween-40, respectively, which are higher than those of other studied surfactants. A high HLB value indicates high hydrophilicity of the surfactant, which may allow good compatibility of the coated MNPs with aqueous sample solution. Hence, TW-20-C12 and TW-40-C12 exhibited better dispersibility and recoveries in aqueous solution. Both of them were chosen for the following study. 3.2. Characterization of TW-20-C12 and TW-40-C12 MNPs As discussed above, TW-20-C12 and TW-40-C12 appeared to have good anti-interference abilities. These two materials were characterized and compared with C12 -Fe3 O4 and Fe3 O4 by FT-IR measurement, as shown in Fig. 2. The absorption peaks around 3440 and 565 cm−1 appeared in all curves corresponded to the –OH and Fe–O group, respectively. In the spectra of C12 -Fe3 O4 , TW-20-C12 and TW-40-C12 , the absorption peaks around 2850 and 2920 cm−1 can be observed, which should be ascribed to the vibration of C–H of the dodecyl group, suggesting that dodecyl group was successfully immobilized onto the MNPs. However, no difference was observed among FT-IR spectra of these three materials.

Fig. 2. FT-IR spectra of (A) Fe3 O4 , (B) C12 -Fe3 O4 , (C) TW-20-C12 and (D) TW-40-C12 .

TGA was carried out to identify the presence of surfactant coating. The weight loss from 473 K to 973 K was 4.85%, 5.36% and 6.10% for C12 -Fe3 O4 , TW-40-C12 and TW-20-C12 , respectively. The data demonstrated that C12 -Fe3 O4 MNPs were successfully coated with the surfactants. Fig. 3 shows TEM images of the C12 -Fe3 O4 , TW-20C12 and TW-40-C12 . It can be found that the size of the material increased in the order of C12 -Fe3 O4 , TW-40-C12 and TW-20-C12 . This observation agreed well with TGA results. After coated with the surfactant, particles became larger. As a reflection, Tween-20 coating resulted in higher carbon content than Tween-40 did.

Fig. 3. TEM images of (A) C12 -Fe3 O4 , (B) TW-20-C12 and (C) TW-40-C12 .

50

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

A HC AD PG TP

2 1 0

5

10

15

20

25

5

3

Peak area (× 10 μV·S)

5

Peak area (× 10 μV·S)

A

3

HC

2

PG TP

0

0

50

4

HC AD PG TP

3 2 1 15

20

25

Amount of TW-40-C 12 (mg) Fig. 4. The influence of surfactant-C12 -MNPs amount on MSPE. Extraction condition: 200 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM, different amounts of MNPs, extraction for 20 min by sonication, desorption in 150 ␮L of MeOH for 10 min.

3.3. MSPE optimization based on TW-20-C12 and TW-40-C12 MNPs 3.3.1. The influence of MNPs amount The sorbent amount is an essential factor for MSPE performance. The influence of surfactant-C12 -MNPs amount on the extraction was studied. As can be seen from Fig. 4(A), peak areas of the analytes increased slightly with the increasing TW-20-C12 amount and the curve flattened over 20 mg. In Fig. 4(B), peak areas changed a little irregularly with the change of TW-40-C12 amount; 5 mg was acceptable for all the test analytes. Therefore, 20 mg TW-20-C12 and 5 mg TW-40-C12 were enough for extraction of target compounds. In addition, TW-40-C12 resulted in stronger analytical signals than TW-20-C12 did, which means that the former one provided better extraction ability towards the test analytes. The reason for this observation may be as follows. Although the two surfactants, Tween-40 and Tween-20, have the equal oxyethylene groups per molecule, Tween-40 have longer hydrocarbon tail than Tween-20. Since hydrocarbon tail provides hydrophobicity, it could ensure surfactant molecules to be well arranged onto the surface of MNPs; it may also make a contribution to extraction when hydrophobic interaction is involved. Hence, TW-40-C12 had better extraction ability than TW-20-C12 . 3.3.2. The influence of extraction time Preliminary experiments indicate that poor reproducibility was obtained when the extraction time was less than 5 min. Therefore, the extraction time in the range of 5–25 min was investigated. Peak areas of ST-HRMs almost remained unchanged with prolonging extraction time over 5 min (data not shown), which indicates that the extraction equilibrium can be achieved in 5 min. It is noticeable that the equilibrium time in the case of C12 -Fe3 O4 was relatively longer (about 30 min). As discussed previously [1–9], nano-sized sorbents possessed high surface area, a short diffusion route and good dispersibility. Owing to the considerable contact between the sorbents and aqueous sample, the equilibrium was attained quickly. However, a little longer time was beneficial to get good reproducibility. Hence, 10 min was chosen for the further study.

Peak area ( × 10 5 μV·S)

B

10

150

200

250

NaCl (mg mL )

5

5

100

-1

B Peak area (× 10 5 μV·S)

AD

1

Amount of TW-20-C12 (mg)

0

TW-20-C12

4

TW-40-C 12 4 HC AD PG TP

3 2 1 0

0

50

100

150

200

250

-1

NaCl (mg mL ) Fig. 5. The influence of salt concentration on MSPE. Extraction condition: Different concentration of NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM, extraction for 10 min by sonication, desorption in 150 ␮L of MeOH for 10 min, (A) 20 mg TW-20-C12 and (B) 5 mg TW-40-C12 .

3.3.3. The influence of desorption time To achieve satisfactory extraction efficiency, the influence of desorption time was investigated. When desorption time increased from 5 to 25 min, all peak areas remained almost unchanged (data not shown). It indicated that the equilibrium state approached in 5 min. In the case of TW-40-C12 , the changes of peak areas of AD, PG and TP were almost indistinctive, which imply that desorption was fast. However, the peak area of HC increased evidently with increasing desorption time from 5 to 10 min. After 10 min, the peak area remained unchanged. Furthermore, concurrent results of all analytes in both systems showed that the experimental precision was enhanced with increasing desorption time. The reason may be that suitable sonication time was helpful to disperse MNPs. Therefore, 10 min of desorption time was chosen based on the experimental results. 3.3.4. The influence of salt addition in the sample solution NaCl was added into the sample solution in the concentration range of 0–250 mg mL−1 to study salt influence. As shown in Fig. 5(A), in the case of TW-20-C12 , with increasing concentration of NaCl from 0 to 100 mg mL−1 , the analytical peak areas increased gradually, and the curves flattened over 100 mg mL−1 . Fig. 5(B) showed indistinctive changes of PG and TP peak areas with increasing salt concentration. However, the obvious stronger HC and AD signals were observed with increasing salt concentration. The optimal results were obtained at 200 mg mL−1 of NaCl in the case of TW-40-C12 . This indicates that salting-out effect (increasing the extraction efficiency) was obvious in both cases. Therefore, 100 and 200 mg mL−1 of NaCl represented a suitable concentration for TW-20-C12 and TW-40-C12 , respectively. 3.3.5. The influence of MeOH addition in the sample solution In our preliminary experiment, addition of MeOH in the sample solution was necessary to disperse C12 -Fe3 O4 and ensured good extraction efficiency. To confirm the role of MeOH when surfactant coated C12 -Fe3 O4 were used, the influence of MeOH addition

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

A

4 HC AD PG TP

3 2 1 0

0

5

10

15

20

5

TW- 20- C12

Peak area (× 10 μV·S)

Peak area (× 10 5 μV·S)

A

TW-20-C12 4 3

1 0

2 1 10

15

20

MeOH% (v/v) Fig. 6. The influence of MeOH addition in the sample solution on MSPE. Extraction condition: Extraction for 10 min by sonication, desorption in 150 ␮L of MeOH for 10 min, (A) 20 mg TW-20-C12 and 100 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM and (B) 5 mg TW-40-C12 and 200 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM.

was studied. As shown in Fig. 6, the peak areas of all analytes decreased evidently in both systems when the volume ratio of MeOH increased from 0 to 20%, which implies that the addition of MeOH had negative effect on extraction efficiency. Since TW-20C12 and TW-40-C12 had good dispersibility in aqueous samples, the addition of MeOH weakened the interaction between C12 groups and the analytes. Therefore, no MeOH was added to the sample solution when surfactant coated-C12 -MNPs were used as sorbents. 3.3.6. The influence of sample pH The pH value is important as it affects the ionization status as well as solubility of the analytes. It may also affect the performance of the sorbents since the charge density of mineral oxide surface varies with pH of the solution. In this study, the influence of varying pH values of the sample solution (in the range of 2.0–8.0) on the extraction was studied. With increasing pH from 2.0 to 8.0, slight changes were observed, as shown in Fig. 7. The probable reason may be that the target compounds are neutral (AD, PG and TP) or weak acidic (HC). The pH has no obvious effect on hydrophobic interaction between these compounds and the sorbents. Furthermore, the results demonstrated that the surface coating had good chemical stability in the test pH range. From previous reports, it is known that pH of sample solution could affect the physical adsorption of surfactants on the MNPs. Li et al. [21] found that mixed hemimicelles MNPs based on CTAB coating exhibited no obvious adsorption when pH was between 5.2 and 6.3. It was explained that CTAB molecules were hardly adsorbed onto the positively charged surface of MNPs when the pH was below its zero charge point (6.5). Afkhami et al. [3] studied the adsorption yield of SDS-coated MNPs in the pH range of 2.0–7.2. Results indicated that the adsorption of SDS onto surface sites decreased significantly at higher pHs (pH > 5). As well, similar results were obtained by Faraji et al. [17]. From these studies, it can be seen that materials based on ionic surfactant coating merely exhibited good stabilities and satisfactory extraction efficiencies in certain pH range, while materials developed in our study could be used in a wider pH scope.

5

Peak area (× 10 μV·S)

5

Peak area (× 10 μV·S)

HC AD PG TP

3

5

3

4

B

4

0

2

5

6

7

8

pH

TW- 40- C12

0

HC AD PG TP

2

MeOH% (v/v)

B

51

TW-40-C 12 5 4

HC

3

AD PG

2

TP

1 0

2

3

4

5

6

7

8

pH Fig. 7. The influence of sample pH on MSPE. Extraction condition: Extraction for 10 min by sonication, different pH of sample solution, desorption in 150 ␮L of MeOH for 10 min, (A) 20 mg TW-20-C12 and 100 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM and (B) 5 mg TW-40-C12 and 200 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM.

3.3.7. The influence of the desorption solvent Acetone, ACN, MeOH and n-hexane are commonly used organic solvents to elute the analytes from the sorbents. Under the same extraction conditions, the interfering peaks appeared at the same elution windows as HC and AD in the case of acetone as an eluant. When n-hexane was used, HPLC signals of analytes were not obvious. Satisfactory results were obtained when ACN and MeOH were used instead. ACN exhibited better desorption of target molecules in the case of TW-40-C12 , while MeOH did better in TW-20-C12 . Thus, ACN and MeOH were selected as eluants for TW-40-C12 and TW-20-C12 , respectively. Ulteriorly, the effect of eluant volume was evaluated over the range of 150–240 ␮L. As shown in Fig. 8, with the increasing eluant volume, the corresponding peak area decreased. Expectedly, lower eluant volume should afford higher enrichment as long as it was enough to elute the analytes. Therefore, 150 ␮L eluant was selected for desorption. In summary, for TW-20-C12 , the optimized extraction conditions were 20 mg of the sorbents, 4 mL sample solution with the addition of 100 mg mL−1 NaCl, extraction and desorption time of 10 min, and 150 ␮L MeOH as an eluant. For TW-40-C12 , the optimized extraction conditions were 5 mg of the sorbents, 4 mL sample solution with the addition of 200 mg mL−1 NaCl, extraction and desorption time of 10 min, and 150 ␮L ACN as an eluant. The conditions were used afterwards. 3.4. Method evaluation All the validation data of respective optimal condition for two sorbents are presented in Table 2. The linearities were investigated over a concentration range of 5/50/100–1000 ng mL−1 and 1/5/20–1000 ng mL−1 for TW-20-C12 and TW-40-C12 , respectively, which were calculated by plotting corresponding HPLC peak areas (y) versus concentrations of studied analytes (x, ng mL−1 ). Intra-day and inter-day relative standard deviations (RSDs) were both examined at three concentration levels, and each level was repeated

1

3.4 3.8 9.5 12.8 5.3 4.1 8.0 5.5 9.8 7.8 1.7 0.6 11.1 6.7 2.1 1.6 0.62 0.19 0.24 0.39 0.9997 0.9984 0.9994 0.9999 20–1000 5–1000 1–1000 5–1000 14.4 7.6 1.5 2.2 8.9 3.1 3.2 7.7 10.1 3.5 3.1 12.0 3.5 4.9 2.1 4.5 13.6 1.5 2.9 5.4

0.8 ␮g mL−1

TW-40-C12

0.9983 0.9943 0.9996 0.9995 100–1500 50–1000 5–1000 5–1000

LOD (ng mL−1 ) TW-20-C12

The proposed method was applied to various matrices under the optimum conditions, including environmental waters (Changjiang River, Hanjiang River and Tap water) and biological fluid (human urine). Although the sample matrices are complex, the only sample preparation procedure required before extraction was to dilute human urine samples with pure water (1:10, v:v). For environmental water samples, no extra sample preparation step was necessary before extraction. As target ST-HMRs were not detected in these samples, spiked real samples were used. The performance of TW20-C12 and TW-40-C12 in spiked samples is illustrated in Table 3. For environmental matrices, reproducibilities and relative recoveries of all test analytes in the case of two materials were acceptable,

Analytes

3.5. Method applications

Table 2 Regression data and LODs of ST-HRMs extracted with TW-20-C12 and TW-40-C12 .

thrice. Limits of detection (LODs) were calculated at a signal-tonoise of 3. In the case of TW-20-C12 , a good linearity was obtained with r2 > 0.9943. Intra-day and inter-day RSDs were 1.5–13.6% and 1.5–14.4%, respectively. LODs for the test compounds were ranging from 0.28 to 20.16 ng mL−1 . In the case of TW-40-C12 , analytes also exhibited good linearity and acceptable intra-day and interday RSDs. LODs were as low as 0.62 (HC), 0.19 (AD), 0.24 (PG) and 0.39 (TP) ng mL−1 , respectively. All the results reveal that the proposed method is suitable to detect ST-HRMs in real environmental and biological samples since these compounds are normally found in ␮g mL−1 concentration levels [24–28]. Additionally, based on the above results, TW-40-C12 provided lower LODs than TW-20-C12 did. Furthermore, the reproducibility of material preparation was studied. Sorbents from three different batches were used to extract analytes from spiked water samples (0.5 ␮g mL−1 ) under the same conditions, and the extraction was repeated in triplicate using individual sorbent. The RSDs of 3.1–17.4% and 1.9–4.9% were obtained for TW-20-C12 and TW-40-C12 , respectively. Acceptable reproducibility indicated that this novel material was promising for real applications.

r2

Fig. 8. The influence of the eluant on MSPE. Extraction condition: Extraction for 10 min by sonication, (A) 20 mg TW-20-C12 , 100 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM and desorption in different volumes of MeOH for 10 min and (B) 5 mg TW-40-C12 , 200 mg mL−1 NaCl in sample solution containing 0.5 ␮g mL−1 of each ST-HRM and desorption in different volumes of ACN for 10 min.

7.8 4.4 5.3 2.5

240

0.8 ␮g mL−1

210

ACN ( μ L)

Linear range (ng mL−1 )

180

Inter-day RSD (%, n = 3)

0 150

LOD (ng mL−1 )

1

20.16 0.82 0.28 0.53

2

Intra-day RSD (%, n = 3)

3

HC AD PG TP

HC AD PG TP

4

r2

5

0.1 ␮g mL−1

Intra-day RSD (%, n = 3)

TW-40-C 12

Linear range (ng mL−1 )

Peak area (× 10 5 μV·S)

B

0.05 ␮g mL−1

240

0.5 ␮g mL−1

210

MeOH (μL)

0.1 ␮g mL−1

180

0.4 ␮g mL−1

0 150

13.1 6.2 4.6 3.5

2

14.2 4.3 2.5 12.7

HC AD PG TP

0.1 ␮g mL−1

3

Inter-day RSD (%, n = 3)

4

0.5 ␮g mL−1

TW-20-C12

0.1 ␮g mL−1

5

Peak area (× 10 μV·S)

A

0.05 ␮g mL−1

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

0.4 ␮g mL−1

52

20 2.6 10.1 8.1 4.7 5.3 2.5 11.6 1.6 3.0 7.1 6.3 10.6 18.6 5.8 11.9 3.5 5.7 5.3 5.4 6.8

0.05 ␮g mL−1

53

4

3

18 14

16.3 16.3 5.1 5.1 8.9 7.7 3.5 3.4 5.9 3.6 3.9 5.8 6.0 1.3 3.0 5.0 7.1 3.0 7.7 14.9

2

12

mAU

3.3 0.9 6.0 10.9 18.0 13.7 7.1 19.3 10.7 11.6 2.2 5.7 8.6 12.4 6.6 7.2 1.8 8.3 5.6 4.2

0.1 ␮g mL−1

16

0.5 ␮g mL−1

Inter-day RSD (%, n = 3)

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

1 Urine sample

10 8

8.7 7.2 1.9 2.5 7.7 8.3 7.0 12.8 5.0 15.3 3.4 9.0 10.0 6.0 1.4 5.3 4.2 5.7 2.8 4.3 11.4 12.1 1.9 4.4 11.6 9.2 4.9 8.8 3.6 7.7 1.3 6.5 9.4 3.7 2.3 3.4 10.6 7.0 10.6 6.5 102.1 95.9 81.2 96.8 115.4 107.9 92.9 107.7 98.2 104.4 98.0 104.2 15.3 21.2 85.9 112.2 26.9 51.7 103.0 120.9 91.2 83.4 99.7 110.1 91.4 84.0 95.4 103.5 90.3 80.8 86.5 100.2 86.8 82.5 98.1 116 92.3 98.1 102.5 112.0 92.7 81.0 103.0 109.4 107.0 95.1 93.9 105.7 114.8 103.3 102.0 113.1 89.8 87.1 101.2 109.2 94.4 87.9 101.2 112.7

4.0 3.2 4.5 4.7 8.0 6.8 3.6 2.8 4.9 1.8 2.9 1.6 7.6 6.2 2.6 1.1 4.4 5.4 0.6 0.2

9.7 6.9 7.4 2.6 13.2 8.7 3.3 1.0 8.8 13.5 11.1 1.0 4.3 4.3 4.8 5.7 6.8 6.7 4.8 11.1

2.4 15.3 11.2 1.3 4.1 4.6 6.9 3.8 8.0 4.6 6.9 3.8 6.8 11.0 1.8 6.4 8.6 5.4 7.1 0.6

13.5 17.1 13.1 8.3 8.7 7.4 9.8 8.4 16.8 13.4 5.7 11.1 12.2 9.9 18.0 5.6 7.6 7.6 0.7 18.5

6.9 8.1 19.4 4.9 14.2 17.7 8.8 10.2 8.1 2.4 7.2 8.6 3.9 19.3 3.9 5.9 6.3 19.5 6.4 8.0

8.0 9.3 6.1 3.0 3.5 8.1 8.9 2.8 7.8 5.6 5.0 1.4 10.7 18.5 8.6 3.5 19.0 15.6 5.4 8.2

94.0 88.1 105.7 108.2 97.4 86.6 94.6 99.4 88.5 98.9 102.5 99.0 28.1 43.7 100.5 122.1 35.9 52.3 98.4 122.9

108.0 95.1 93.6 101.4 99.8 92.0 94.4 91.3 104.4 96.7 93.4 88.2 25.8 39.7 99.5 139.9 29.4 48.1 98.9 118.9

13.9 13.2 6.4 5.0 13.1 11.9 2.4 1.9 6.9 2.2 1.4 2.2 5.9 8.0 16.4 4.3 1.0 14.5 9.8 5.7

0.05 ␮g mL−1 0.1 ␮g mL−1 0.5 ␮g mL−1 0.05 ␮g mL−1 0.1 ␮g mL−1 0.5 ␮g mL−1 0.1 ␮g mL−1 0.4 ␮g mL−1 0.8 ␮g mL−1 0.1 ␮g mL−1 0.4 ␮g mL−1 0.8 ␮g mL−1 0.1 ␮g mL−1 0.4 ␮g mL−1

Human urine (female)

Human urine (male)

Tap water

Hanjiang River

HC AD PG TP HC AD PG TP HC AD PG TP HC AD PG TP HC AD PG TP Changjiang River

115.4 106.7 108.7 113.1 110.6 104.3 99.3 102.4 118.5 109.6 105.6 105.6 95.8 96.6 101.9 113.8 109.9 101.2 97.9 102.5

Relative recovery (%)

Ultrapure water

-2 0

2

4

6

8

10

12

14

16

18

20

Time (min)

Analytes

Intra-day RSD (%, n = 3)

2

Fig. 9. HPLC–UV chromatograms of different samples extracted by TW-20-C12 : (A) ultrapure water, (B) Hanjiang River water and (C) urine sample. All of these samples were spiked with 0.5 ␮g mL−1 of each ST-HRM. Peak assignment: (1) HC, (2) AD, (3) PG and (4) TP.

which indicate that both Tween-40 and 20 coatings played an antiinterference role, and can be used as RAM coating for extraction. However, for urine samples, the relative recoveries of HC and AD in the case of TW-40-C12 were much lower than those in TW-20-C12 . This observation may be due to the fact that biological sample was more complex than environmental matrices. For complex sample matrices, owing to its high HLB value, Tween-20 coating may have better biological compatibility than Tween-40, thus resulting in better anti-interference ability. Fig. 9 shows HPLC chromatograms when TW-20-C12 MNPs were used as sorbents. As discussed above, this material was promising for extraction of complex samples. Various sample preparation methods of ST-HRMs have been applied prior to the instrumental analysis, such as solid-phase disk extraction [24], vial wall sorptive extraction [29], liquid–liquid extraction [30], dispersive liquid–liquid microextraction [31] and SPE [25,32]. These methods are predominantly used for environmental samples. Some of them were associated with multi-steps, which made extraction operation tedious. The present developed approach is fast, convenient and only one-step extraction is involved. Additionally, the method afforded comparable or even better LODs and RSDs compared to previous work [24–26,28–32], which was satisfactory for monitoring ST-HRMs in environmental and biological samples. Furthermore, it was demonstrated that RAM-MSPE based on non-ionic surfactant coatings had potential applications in the complex sample matrices with anti-interference ability. It could be complementary to the present available methods for the determination of ST-HRMs.

4. Conclusions 0.8 ␮g mL−1

TW-40-C12

Relative recovery (%)

Inter-day RSD (%, n = 3) TW-20-C12

Hanjiang River water

4 0

Genuine samples

8 Table 3 Relative recoveries and RSDs of ST-HRMs in spiked real samples extracted with TW-20-C12 and TW-40-C12 .

Intra-day RSD (%, n = 3)

6

In this study, RAM-C12 -Fe3 O4 magnetic nanoparticles (MNPs) based on non-ionic surfactant coating were successfully prepared and used to extract steroid hormones from environmental and biological samples. The material combines the advantages of RAM and MNPs, affording good anti-interference ability, good dispersibility in aqueous solution, fast extraction and convenient operation. They were successfully applied to several environmental and biological samples. Satisfactory results including acceptable recovery, reasonable reproducibilities and low LODs were achieved. This novel material is promising as an anti-interference sorbent for applications in complex samples.

54

L. Ye et al. / J. Chromatogr. A 1244 (2012) 46–54

Acknowledgments The authors gratefully acknowledge the financial support of this research by the Nature Science Foundation of China (Nos. 81173018, 21105033 and 21177099), the Fundamental Research Funds for the Central Universities (Nos. 01-09-514944 and 114027) and New Teachers’ Fund for Doctor Stations, Ministry of Education (No. 20110142120029). References [1] C. Huang, B. Hu, J. Sep. Sci. 31 (2008) 760. [2] Y.S. Ji, X.Y. Liu, M. Guan, C.D. Zha, H.Y. Huang, H.X. Zhang, C.M. Wang, J. Sep. Sci. 32 (2009) 2139. [3] A. Afkhami, R. Moosavi, T. Madrakian, Talanta 82 (2010) 785. [4] H.H. Yang, S.Q. Zhang, X.L. Chen, Z.X. Zhuang, J.G. Xu, X.R. Wang, Anal. Chem. 76 (2004) 1316. [5] X.L. Pu, Z.C. Jiang, B. Hu, H.B. Wang, J. Anal. Atom. Spectrom. 19 (2004) 984. [6] X.L. Zhang, H.Y. Niu, S.X. Zhang, Y.Q. Cai, Anal. Bioanal. Chem. 397 (2010) 791. [7] X.L. Zhao, Y.L. Shi, T. Wang, Y.Q. Cai, G.B. Jiang, J. Chromatogr. A 1188 (2008) 140. [8] X.L. Zhang, H.Y. Niu, Y.Y. Pan, Y.L. Shi, Y.Q. Cai, Anal. Chem. 82 (2010) 2363. [9] L.G. Chen, T. Wang, J. Tong, Trends Anal. Chem. 30 (2011) 1095. [10] L. Novakova, H. Vlckova, Anal. Chim. Acta 656 (2009) 8. [11] P. Sadilek, D. Satinsky, P. Solich, Trends Anal. Chem. 26 (2007) 375. [12] C.L. Wa, R. Mallik, D.S. Hage, Anal. Chem. 80 (2008) 8751.

[13] K. Hoshina, S. Horiyama, H. Matsunaga, J. Haginaka, J. Pharm. Biomed. Anal. 55 (2011) 916. [14] Z. Li, D. Huang, C. Fu, B. Wei, W. Yu, C. Deng, X. Zhang, J. Chromatogr. A 1218 (2011) 6232. [15] X.-L. Zhang, H.-Y. Niu, W.-H. Li, Y.-L. Shi, Y.-Q. Cai, Chem. Commun. 47 (2011) 4454. [16] Y. Wang, Y. Wang, L. Chen, Q.-H. Wan, J. Magn. Magn. Mater. 324 (2012) 410. [17] M. Faraji, Y. Yamini, M. Rezaee, Talanta 81 (2010) 831. [18] C.P. Desilets, M.A. Rounds, F.E. Regnier, J. Chromatogr. 544 (1991) 25. [19] J.P. Chang, J. Chromatogr. A 317 (1984) 157. [20] Z.F. Wang, H.S. Guo, Y.L. Yu, N.Y. He, J. Magn. Magn. Mater. 302 (2006) 397. [21] J.D. Li, X.L. Zhao, Y.L. Shi, Y.Q. Cai, S.F. Mou, G.B. Jiang, J. Chromatogr. A 1180 (2008) 24. [22] T. Saitoh, Y. Nakayama, M. Hiraide, J. Chromatogr. A 972 (2002) 205. [23] L. Sun, C. Zhang, L. Chen, J. Liua, H. Jin, H. Xu, L. Ding, Anal. Chim. Acta 638 (2009) 162. [24] L. Sun, Y. Liu, X.G. Chu, J.M. Lin, Chromatographia 71 (2010) 867. [25] T. Trinh, N.B. Harden, H.M. Coleman, S.J. Khan, J. Chromatogr. A 1218 (2011) 1668. [26] M.G. Zheng, L. Wang, Y.D. Bi, F. Liu, J. Environ. Sci. 23 (2011) 693. [27] S. Gorog, J. Pharm. Biomed. Anal. 55 (2011) 728. [28] S. Flor, S. Lucangioli, M. Contin, V. Tripodi, Electrophoresis 31 (2010) 3305. [29] M. Kawaguchi, S.-I. Fujii, N. Itoh, R. Ito, H. Nakazawa, A. Takatsu, J. Chromatogr. A 1216 (2009) 7553. [30] M.A. Soliman, J.A. Pedersen, I.H. Suffet, J. Chromatogr. A 1029 (2004) 223. [31] X.W. Du, X.D. Wang, Y.Y. Li, F.Q. Ye, Q.X. Dong, C.J. Huang, Chromatographia 71 (2010) 405. [32] E. Vulliet, J.B. Baugros, M.M. Flament-Waton, M.F. Grenier-Loustalot, Anal. Bioanal. Chem. 387 (2007) 2143.