- Email: [email protected]
Dynamic headspace time-extended helix liquid-phase microextraction
Journal of Chromatography A, 1216 (2009) 4347–4353
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Dynamic headspace time-extended helix liquid-phase microextraction Shih-Pin Huang, Pai-Shan Chen, Shang-Da Huang ∗ Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan
a r t i c l e
i n f o
Article history: Received 11 November 2008 Received in revised form 9 February 2009 Accepted 13 March 2009 Available online 18 March 2009 Keywords: Headspace LPME Dynamic hollow fiber-LPME Organochlorine Pesticides GC–MS
a b s t r a c t Liquid-phase microextraction (LPME) has been proved to be a fast, inexpensive and effective sample pre-treatment technique for the analyses of pesticides and many other compounds. In this investigation, a new headspace microextraction technique, dynamic headspace time-extended helix liquid-phase microextraction (DHS-TEH-LPME), is presented. In this work, use of a solvent cooling system, permits the temperature of the extraction solvent to be lowered. Lowering the temperature of the extraction solvent not only reduces solvent loss but also extends the feasible extraction time, thereby improving extraction efficiency. Use of a larger volume of the solvent not only extends the feasible extraction time but also, after extraction, leaves a larger volume to be directly injected into the gas chromatography (GC) to increase extraction efficiency and instrument signal. The DHS-TEH-LPME technique was used to extract six organochlorine pesticides (OCPs) from 110 ml water samples that had been spiked with the analytes at ng/l levels, and stirred for 60 min. The proposed method attained enrichments up to 2121 fold. The effects of extraction solvent identity, sample agitation, extraction time, extraction temperature, and salt concentration on extraction performance were also investigated. The method detection limits (MDLs) varied from 0.2 to 25 ng/l. The calibration curves were linear for at least 2 orders of magnitude with R2 0.996. Relative recoveries in river water were more than 86%. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Organochlorine pesticides (OCPs) were introduced after World War II for the control of vector-borne diseases and were also used as insecticides in agriculture [1]. However, OCPs are known to accumulate in the tissues of living organisms, and their general worldwide extensive usage has led to substantial bioaccumulation. These compounds pose potential hazards to human health and the environment. OCPs have now been banned in most developed countries. However, the early spectacular success of dichlorodiphenyltrichloroethane (DDT) for malaria eradication in some countries has resulted in this compound’s continued use in developing countries. Studies have suggested that DDT and related compounds may affect the normal function of the endocrine system [2,28]. The ability of the prevalent isomer of the major and most persistent DDT derivative, p,p -dichlorodiphenyldichloroethylene (p,p -DDE), to bind to the androgen receptor in male rats has been reported [3,4]. For environmental and drinking water, the European Union (EU) has established the maximum admissible concentration of a single compound as 0.1 g/l, and 0.5 g/l is the maximum allowed for the total concentration of all OCPs [5,6]. These very low
∗ Corresponding author. Tel.: +886 3 572 1194; fax: +886 3 573 6979. E-mail address: [email protected] (S.-D. Huang). 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.03.030
concentrations make the development of an efficient and sensitive analytical method necessary. The detection of such contaminants is an important task for chemists working in the field of environment protection. Analyte extraction and pre-treatment is the most challenging and time-consuming step in an analytical procedure. Many extraction procedures are available viz. liquid–liquid extraction (LLE) and solid-phase extraction (SPE) techniques [29]. However, LLE is time consuming, generally labor intensive, and requires large quantities of expensive, toxic and environmentally unfriendly organic solvents. The conventional offline SPE technique requires less solvent, but is relatively expensive. In analytical chemistry, the trend is toward simplification and miniaturization of the sample preparation steps, and toward a decrease in the quantities of organic solvents used. More recently, a novel technique, solid-phase microextraction (SPME) [30–32], solvent microextraction (SME) [7,8] or single-drop microextraction (SDME)/liquid-phase microextraction (LPME) [9,10], has been developed. In 2002, Shen and Lee developed a LPME method based on a protected hollow-fiber membrane combined with the microsyringe and termed it hollow-fiber liquid-phase microextraction (HF-LPME) [11,12]. LPME has been shown to be an advantageous alternative sample preparation method to conventional LLE, and is relatively low in cost compared with SPE and SPME. LPME requires the use of quite small volumes (e.g., 5 l or less) of organic solvent to extract analytes from aqueous matrices.
4348
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Recently, two major variants have been developed for hollowfiber-protected LPME: direct-immersion and headspace. The latter, an analytical method for volatile organic compounds (VOCs) in aqueous or other matrices, is utilized widely as a method of directly detecting VOCs without the interference of a dirty matrix that could occur in the direct-immersion mode [13,14]. In such analysis, a more efficient sampling method is required if one is dealing with semivolatile compounds or with VOCs present at trace levels. However, the main problem of the technique is that all the commonly used organic solvents for GC analysis have high vapor pressures, which causes them to evaporate too rapidly in air. Moreover, when using water-miscible solvents in single drop microextraction, the increase in drop size that occurs during sampling may cause the drop to fall from the needle [15]. Apparently, the selection of suitable organic solvents is limited, as the vapor pressure of the solvent must be considered in addition to its extraction efficiency. Selection of a solvent which is compatible with GC and suitable for analysis in the headspace is a key step. Theis et al. [16] reported headspace single drop microextraction in which 1-octanol, a solvent with a very low vapor pressure, was used. Shen and Lee [17] have introduced dynamic headspace LPME to overcome some limitations of static headspace single drop microextraction. Jiang et al. [18] developed a dynamic hollow-fiber-supported headspace liquid-phase microextraction (DHF-HS-LPME) approach. A hollow fiber filled with the extracting solvent is placed onto a microsyringe needle and placed in the headspace of the sample solution. The solvent within the fiber is moved to-and-fro by means of a programmable syringe pump. However, there are some problems with these techniques. In DHF-HS-LPME, the extracting solvent within the straight fiber (1.5 cm) will stack and gradually form a droplet at the end of the fiber during extraction. This causes air bubbles to enter the fiber. Because of this phenomenon, it is difficult for the syringe pump to withdraw extracting solvent or to just withdraw discontinuous portions of solvent into a microsyringe after the extraction procedure. [19] Therefore, longer fibers which possess more extraction contact interface and solvent volume may not be suitable for use in headspace analysis. On the other hand, for headspace LPME, heating specimen solution apparently enhances the extraction efficiency, but it also hastens the evaporation of the limited amount of extracting solvent. Also, the solvent volume of a single drop microextraction or a DHF-HS-LPME cannot be too large. To address the above-mentioned problems, an efficient headspace LPME technique, dynamic headspace time-extended helix liquid-phase microextraction (DHS-TEH-LPME), is developed. Here, a long fiber (4.5 cm) is bent as a helix shape (Fig. 1). One end of the fiber is affixed to a pin and the other end is inserted into a microsyringe which contains 13 l of extracting solvent. The solvent is moved back and forth in the fiber by a programmable syringe pump. The exceed solvent that is held outside of the fiber forms a large droplet that is suspended at the ring portion of a helix-shaped fiber. As for the shape of the helix-type fiber, the extracting solvent does not stack at the end of the fiber as in the case of DHS-TEHLPME. Without interference of air bubbles, the solvent maintaining an unbroken form is withdrawn into the microsyringe easily and completely. Also, the long fiber and the large droplet provide a larger contact surface and larger volume of organic solvent. In DHS-TEHLPME analysis, the extraction time can be extended in the following two ways. In one of them, the solvent cooling assisting system, lowering the temperature of the extracting solvent reduces solvent loss. The other way is by increasing the initial volume of extracting solvent. Because solvent loss is reduced, this factor may be ignored during the dynamic headspace microextraction. Therefore, after extraction, a larger volume is available to be directly injected into the gas chromatography (GC). This improves extraction efficiency and signal amplitude.
Fig. 1. Experimental setup of SC-DHS-TEH-LPME.
Analytes are extracted into a microliter volume of extracting solvent, so analytes may potentially be enriched by factors of hundreds or thousands, depending on their distribution constants and the sample volume [20]. In this work, the large sample volumes (110 ml) used were intended to simulate ultra-trace level determinations of OCPs in river waters. The proposed method could attain enrichments of up to 2121-fold. In order to achieve the best extraction efficiency, the influence of parameters such as organic solvent, temperature, stirring rate, extraction time and ionic strength was investigated. The optimized DHS-TEH-LPME technique was then successfully applied to determine OCPs in water samples. Detection was performed by GC–MS.
2. Experimental 2.1. Reagents and materials All chemicals were of reagent grade and were used without further purification. Heptachlor, aldrin, endosulfan, p,p -DDE, dieldrin, and o,p-DDT were purchased from Fluka (Buchs, Switzerland). Stock standard solutions of the studied pesticides were prepared in methanol at a concentration of 0.1 mg/ml and stored at 4 ◦ C. Working solutions used to optimize the parameters in LPME were prepared at a concentration of 10 ng/ml every day. Purified water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). Sodium chloride was bought from Showa Chemicals (Tokyo, Japan). The extracting solvents 1-octanol, 1hexanol, 1-pentanol, nonane and o-xylene were purchased from Fluka (analytical-grade). A Q 3/2 Accurel Polypropylene hollow-fiber membrane (600 m I.D., 200 m wall thickness, 0.2 m pore size) was purchased from Membrana GmbH (Wuppertal, Germany). The hollow fiber was cut into 4.5 cm segments and was cleaned in acetone and dried before use.
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
4349
2.2. Instrumentation Analyses were performed using a Varian (Walnut Creek, CA, USA) CP-3800 gas chromatography system with a 1079 injector kept at 260 ◦ C, in splitless mode (1 min). A ChromatoProbe (Varian Instruments, Walnut Creek, CA, USA) was coupled with this injector. A 30m DB-5MS (0.25 mm I.D., 0.25 mm film thickness) fused-silica capillary column (J&W Scientific, Folsom, CA, USA) was used for all analyses; helium at a constant flow of 1 ml/min was the carrier gas. The temperature was held at 150 ◦ C for 2 min, increased to 230 ◦ C at 20 ◦ C/min and held for 2 min, then raised to 280 ◦ C at 20 ◦ C/min and held for 6 min. A Varian Saturn 2000 ion trap GC–MS system was used for tandem mass spectrometric (MS–MS) detection. The ion trap mass spectrometer was operated in the electron ionization (EI) mode; the ionization energy was set at 70 eV and the source temperature at 200 ◦ C. Full-scan MS–MS data were acquired over the m/z range 100–450 to obtain the fragmentation spectra of the analytes. The Varian-supplied routine, ‘Automated Methods Development (AMD)’, was used to find the optimum collision-induced dissociation (CID) excitation amplitude. Specific product ions in the MS–MS spectra were used for detection in this study, and the appropriate extracted ion chromatograms were calculated and integrated for quantitation. The qualitative GC–MS–MS data for the pesticides studied are shown in Table 1. 2.3. Extraction process A 3 cm × 0.6 cm stir bar was put into 110 ml of sample solution in a 120 ml home made vial. The syringe was mounted in a cooling assisting system and was fitted on the syringe pump. A headspace sample vial septum cap was pierced by the syringe. A 13 l portion of 1-octanol was drawn into the microsyringe. The needle tip was inserted into the 4.5 cm hollow fiber. The hollow fiber was revolved around the syringe needle, which formed in a helix shape. The end of the fiber was fixed on a pin, as shown in Fig. 1. Subsequently the assembly was then immersed in 1-octanol to impregnate the pores of the hollow fiber for 15 s to impregnate the pores of the fiber. Then the prepared fiber was removed and put in the headspace above a 110 ml aqueous sample which contained 10 g/l of the target compounds. The syringe pump was pre-programmed to dispose and withdraw 10 l of the organic phase in a continuous mode at a speed of 40 l/min and the cooling system was set at −10 ◦ C. The solution was then stirred at 600 rpm at 50 ◦ C for 30 min. The final movement of the plunger was to withdraw the extract into the syringe barrel and injected 10 l into the GC–MS for analysis. The used fiber was discarded and a fresh one was used for the next experiment. 3. Results and discussion 3.1. Selection of organic solvent Selection of a suitable organic extraction solvent is crucial for dynamic headspace LPME. The selection criteria for an organic solvent to be a suitable choice are that (1) it should be easily immo-
Fig. 2. Effect of extraction solvent on LPME efficiency (n = 3), spiked to 10 g/l of each of the analytes, 600 rpm, 50 ◦ C, 30 min.
bilized in the pores of the polypropylene hollow fiber, and (2) the organic solvent should be of low volatility to prevent solvent loss, (3) volatile enough for use in GC–MS analysis. In consideration of these factors, o-xylene, nonane, 1-pentanol, 1-hexanol and 1octanol were tested for their suitability. Fig. 2 exhibits the extraction efficiencies for the tested organic solvents. According to Fig. 2, the extraction efficiency was best when 1-octanol was used for most OCPs except p,p -DDE. The reason could be that the 1-octanol, 1hexanol and 1-pentanol are more polar than o-xylene and nonane and high polarity enhances better dissolution because OCPs are polar compounds. However, the volatility of 1-octanol is lower than that of 1-hexanol and 1-pentanol. The lower volatility the solvent has, the less volume the solvent will lose the longer the extraction time will take in the process. Therefore the selection of organic solvent is more flexible for this approach than is the case for the previous headspace LPME method. Hence for further experiments 1-octanol was chosen as the organic extractant. 3.2. Effect of stirring speed HS-LPME is based on equilibrium partitioning of the analyte between three phases: the condensed aqueous phase, the headspace phase and the organic solvent phase. Mass transfer in the headspace is assumed to be a faster process than it is in either of the condensed phases since diffusion coefficients in the gas phase are typically about 104 times greater than corresponding diffusion coefficients in the condensed phase. Thus, both slow mass transfers in the aqueous phase and diffusion of analyte into the organic solvent phase are limiting steps in the overall extraction process [21]. The aqueous phase mass transfer coefficient of solute increases with increasing stirring rate. Therefore, agitation of the sample solution improves mass transfer in the aqueous phase and induces convection in the headspace, so consequently the equilibrium between the aqueous and vapor phase can be attained more rapidly. Fig. 3 describes the influence of stirring speed on the extraction efficiency. The extraction efficiency increased with increasing stirring rate up
Table 1 Qualitative GC/MS/MS data for the studied pesticides. Analyte
Retention time (min)
Precursor ion (m/z)a
CID excitation amplitude (V)
Product ions, m/z (Relative abundance (%))
Quantitation ions (m/z)
Heptachlor Aldrin Endosulfan p,p -DDE Dieldrin o,p -DDT
6.04 6.57 7.65 7.85 8.02 8.49
272 293 241 318 279 235
0.32 0.33 0.40 0.38 0.40 0.47
272(12), 270(100), 237(79) 293(25), 291(100), 257(56) 241(27), 239(100), 206(45) 318(37), 316(100), 283(17) 279(28), 277(100), 243(34) 235(21), 200(100), 165(39)
270/237 291/257 239/206 316/283 277/243 200/165
a
Not the molecular ion, selected on basis of signal intensity.
4350
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Fig. 3. Effect of stir rate on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 50 ◦ C, 30 min.
Fig. 5. Effect of temperature on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 60 min, stirring at 800 rpm.
to 800 rpm. However, when the stirring rate exceeded 800 rpm, the stirring bar in the sample solution could not move steadily, and consequently lowered the extraction efficiency. Finally, a stirring speed of 800 rpm was used for subsequent experiments.
sample matrix/headspace/hollow-fiber-supported organic phase. The effect of extraction temperature was evaluated from 50 to 90 ◦ C. As seen in Fig. 5, the extraction efficiency of OCPs increases with increasing extraction temperature. When the sample solution was heated to a temperature higher than 80 ◦ C, extraction efficiency dropped, which may due to the loss of extracting solvent, and the competition between Henry’s constant and the partition coefficient. At higher temperatures, both the Henry’s constants and the diffusion coefficients of the target analytes become larger. This results in an increase of the vapor pressure of the OCPs. Therefore their concentrations also rise in the headspace [23]. However, the partition coefficient between the headspace and the organic phase is reduced with increasing temperature [24]. For these reasons, a temperature of 80 ◦ C was used for further work.
3.3. Effect of extraction time Dynamic headspace LPME is not an exhaustive extraction technique; rather, it is an equilibrium processes in which analytes partition between the condensed phase/headspace, and the headspace/extracting organic solvent. Hence, extraction efficiency depends on the length of the extraction time. Fig. 4 shows the influence of extraction time on efficiency of the OCPs. The peak areas increase with increasing extraction time in the range of 30–65 min. However, longer extraction times (>65 min) may lead to too much water vapor condensation which may freeze in the syringe needle and interfere with efficient extraction. On the other hand, exposure time in the extraction step should not be excessively long, as would be the case if one were trying to reach equilibrium; extraction time should be just long enough to give satisfactory sensitivity and precision [9,22]. Thus the extraction time was set at 60 min in the following experiments.
3.5. Effect of ionic strength
For headspace analysis of semi-volatile compounds, another important parameter is temperature. Temperature has a significant effect on both the kinetics and the thermodynamics of the extraction process. Temperature affects the kinetics of sorption in the extracting organic solvent by determining the vapor pressures of analytes and diffusion coefficient values in the three phases [21]
The effect of ionic strength on extraction efficiency was studied by preparing standard solutions of the analytes together with NaCl at concentrations varying from 0 to 30% (w/v). Adding salt to the aqueous samples may have several effects on the analysis. In generally, a salting out effect has been observed in LLE and solid-phase microextraction (SPME). Salt can decrease the solubilities of analytes in aqueous samples and can enhance their partitioning into the organic phase (LLE) or the adsorbent (SPME). However, the presence of salt was observed to limit the extraction of analytes by some authors [18,25–27]. In our study, as shown in Fig. 6, no increase in extraction was observed after the addition of the sodium chloride. On the contrary, the extraction efficiencies were highest without addition of the sodium chloride, and subsequently decreased as more was added. This is due to the fact that polar molecules may
Fig. 4. Effect of extraction time on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 50 ◦ C.
Fig. 6. Effect of NaCl concentration on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 60 min, 800 rpm, 80 ◦ C.
3.4. Effect of temperature
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Fig. 7. Comparisons of SC-DHS-TEH-LPME performance with DHS-TEH-LPME (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 80 ◦ C. The extraction time of DHS-TEH-LPME was set at 30 min, and SC-DHS-TEHLPME was set at 30 and 60 min.
participate in electrostatic interactions with the salt ions in solution [33], thereby reducing the rate of mass transfer. Hence, further extractions were made without the addition of NaCl. 3.6. Effect of solvent cooling assisting system The solvent cooling assisting system was used to lower the temperature of the organic solvent on the fiber, which prevents extracting solvent from being lost by evaporation at higher temperatures. At too high a temperature (>0 ◦ C), the effects of the cooling system could not be readily observed. At too low a temperature (<−20 ◦ C), the viscosity of the solvent increases, apparently, which might lead to a smaller amount of extraction volume withdrawn by the syringe. Based on these reasons a cooling temperature of −10 ◦ C was used for further studies.
4351
Fig. 8. Comparisons of DHS-TEH-LPME performance with SC-DHS-TEH-LPME (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 80 ◦ C. The extraction time of SC-DHS-TEH-LPME was set at 30 min, and DHS-TEHLPME was set at 30–60 min.
With the solvent cooling assisting system, the temperature of the extraction solvent could be lowered, which lowers the vapor pressure, decreases the loss of solvent, and permits one to extend the extraction time. Comparison between extractions with and without the solvent cooling assisting system is shown in Fig. 7. Without the solvent cooling assisting system, if we want to maintain the GC input amount at 10 l, the longest extraction possible is 30 min. However, with the solvent cooling assisting system, the extraction time can be extended to 60 min. On the other hand, for a given extraction time, it was observed that, as expected, at a lower solvent temperature the extraction yield was increased. The main reason is that the process of analyte absorption in the solvent is exothermic and the partition coefficient is temperature dependent [21]. If a low solvent temperature is maintained during sampling, the partition coefficient is increased. As a result, greater mass transfer of analyte into the solvent occurs and sensitivity significantly increases.
Fig. 9. Chromatogram of OCPs obtained by LPME under optimized conditions (A) unspiked river water; (B) spiked 1.0 g/l in the river water. Peaks: (1) heptachlor; (2) adrin; (3) endosulfan; (4) p,p -DDE; (5) dieldrin; (6) o,p -DDT.
4352
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Table 2 Performance of the validation analysis. Pesticides Heptachlor Aldrin ␣-Endosulfan p,p -DDE Dieldrin o,p -DDT
Linearity (g/l) 0.01–1 0.01–1 0.1–50 0.01–1 0.1–50 0.01–1
R2
MDLsa (ng/l)
RSD (%)b (n = 5)
Absolute Recoveryc (%)
Relative Recoveryd (%)
0.999 0.998 0.997 0.998 0.996 0.997
0.42 0.33 19 0.27 25 0.37
16.3 15.4 12.5 17.8 16.3 17.4
6.2 10.1 5.8 19.3 10.8 13.8
86 106 109 104 106 98
± ± ± ± ± ±
10 17 5 15 12 15
Enrichment Factorb (fold) 677 1112 633 2121 1184 1520
a MDL are calculated as three times the standard deviation of seven replicated runs of standard solution: amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. b Amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. c A percentage value obtained considering spiked amount in the river water as 100%; amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. d A percentage value obtained considering extraction yields in the river water as 100%; amount of analytes added was 1.0 g/l for each analyte; mean ± standard deviation (n = 3).
Table 3 Comparisons of performance of DHS-TEH-LPME and SC-HF-LPME-GC–MS in this study. Pesticides
Heptachlor Aldrin ␣-Endosulfan p,p -DDE Dieldrin o,p -DDT a b
DHS-TEH-LPME
SC-HF-LPME-GC–MS
MDLs (ng/l)
RSDa (%)
Enrichment Factor (fold)
MDLs (ng/l)
RSDb (%)
Enrichment Factor (fold)
0.42 0.33 19 0.27 25 0.37
16.3 15.4 12.5 17.8 16.3 17.4
677 1112 633 2121 1184 1520
42 51 125 53 209 44
14.2 5.5 15.2 9.7 14.1 12.3
141 125 211 65 100 82
Amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. Amount of analytes added was 1.0 g/l for each analyte.
3.7. Comparisons of DHS-TEH-LPME performance with SC-DHF-HS-LPME SC-DHF-HS-LPME was evaluated with the same extraction conditions as DHS-TEH-LPME (expect extraction time) for water samples spiked at a pesticide concentration of 10 g/l; the results are shown in Fig. 8. It could be directly observed from the experiment of SC-DHF-HS-LPME that when the extraction time took more than 30 min, there would be water vapor condensed in the syringe needle. It may lead to the extraction efficiency to decline. Thus the extraction time of SC-DHF-HS-LPME was set at 30 min. At the same extraction time, the peak area of DHS-TEH-LPME was higher than SC-DHF-HS-LPME. It was highly possible reason that DHS-TEHLPME provides a larger contact surface of organic solvent. On the other hand, the extraction time of DHS-TEH-LPME could last longer, so its extraction efficiency is better. 3.8. Quantitative aspects The developed procedure was applied to the determination of OCPs in the river water. A general procedure was followed to extract the target analytes. Fig. 9 shows typical chromatograms obtained from the analysis of the river water. No OCPs was found in the unspiked river water. Repeatability, linearity, correlation coefficient, detection limit and enrichment factors were investigated under optimized experimental conditions. The performance of the developed procedure is summarized in Table 2. It can be seen that the RSD values are smaller than 17.8% based on the peak areas for five replicate runs. This large value is probably due to the fact that large sample volumes (110 ml) were used. Each analyte exhibited a good correlation coefficient, with R2 ≥ 0.996. Relative recoveries, determined as the ratio of the concentration found in the river water and in deionized water samples, spiked with the same amount of analytes were calculated. The relative recoveries of spiked OCPs at 1.0 g/l were between 86 and 109%. The method detection limits (MDLs) values were calculated as three times the standard deviation
of seven replicate runs of river water spiked with low concentrations of the analytes. The MDLs range from 0.27 to 0.42 ng/l, except ␣-endosulfan (19 ng/l) and dieldrin (25 ng/l). The enrichment factors, defined as the ratio between the final (equilibrium) concentration of analyte in the organic phase and the initial concentration of analyte in the aqueous sample, were between 633 and 2121. In comparison with a previously reported headspace liquidphase microextraction method [19] in Table 3, we obtained much lower MDLs and higher enrichment for all targeted analytes. 4. Conclusions In the present study a high-enrichment sample preparation method for the analysis of trace organochlorine pesticides, based on a dynamic headspace time-extended helix liquid-phase microextraction technique has been developed. This method involves the use of only ∼13 l of organic solvent for each extraction. The potential for a dynamic liquid-phase microextraction technique using a commercially available syringe pump has been demonstrated by analyzing OCPs from river water. The method provides extracts with high-enrichment factors along with excellent sample clean-up. Thus the method provides preconcentration and sample clean-up simultaneously. However, the greatest weakness of DHS-THE-LPME was that it could not be fully automated. Besides, due to the greater volume of the sample solution, the time for extraction equilibrium took longer. Thus, it led to higher RSD. Acknowledgment The support of the National Science Council, Taiwan (NSC 962113-M-007-030-MY3) is acknowledged. References [1] H.P. Lia, G.C. Lib, J.F. Jen, J. Chromatogr. A 1012 (2003) 129. [2] E. Hileman, Chem. Eng. News 31 (January) (1994) 19.
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353 [3] W.R. Keice, R.C. Stone, S.C. Law, L.E. Gray, J.A. Kemppainen, E.M. Wilson, Nature 375 (1995) 581. [4] G.R. van der Hoff, P. van Zoonen, J. Chromatogr. A 843 (1999) 301. [5] European Union, Council Directive 98/83/EC on the quality of water intended for human consumption, Off. J. Eur. Commun. 32 (1998) 330. [6] National Primary Drinking Water Regulation (40 CFR Part 141), Fed. Regis. 19 (2002) 426. [7] M.A. Jeannot, F.F. Cantwell, Anal. Chem. 68 (1996) 2236. [8] M. Ma, F.F. Cantwell, Anal. Chem. 70 (1998) 3912. [9] Y. He, H.K. Lee, Anal. Chem. 69 (1997) 4634. [10] Y. Wang, Y.C. Kwok, Y. He, H.K. Lee, Anal. Chem. 70 (1998) 4610. [11] G. Shen, H.K. Lee, Anal. Chem. 74 (2002) 648. [12] L. Zhao, H.K. Lee, Anal. Chem. 74 (2002) 2486. [13] J. Drozd, J. Novak, J. Chromatogr. 165 (1979) 141. [14] A.J. Nunez, L.F. Gonzalez, J. Janak, J. Chromatogr. 300 (1984) 127. [15] A. Tankeviciute, R. Kazlauskas, V. Vickackaite, Analyst 126 (2001) 1674. [16] A.L. Theis, A.J. Waldack, S.M. Hansen, M.A. Jeannot, Anal. Chem. 73 (2001) 5651. [17] G. Shen, H.K. Lee, Anal. Chem. 75 (2003) 98. [18] X.M. Jiang, C. Basheer, H. Zhang, H.K. Lee, J. Chromatogr. A 1087 (2005) 289.
4353
[19] S.P. Huang, S.D. Huang, J. Chromatogr. A 1176 (2007) 19. [20] S.H. Tung, V. Terje, A. Trude, J. Einar, E.R. Knut, P.-B. Stig, Anal. Chim. Acta 592 (2007) 1. [21] A. Przyjazny, J.M. Kokosa, J. Chromatogr. A 977 (2002) 143. [22] M.H. Ma, F.F. Cantwell, Anal. Chem. 71 (1999) 388. [23] M. Saraji, J. Chromatogr. A 1062 (2005) 15. [24] J. Zhang, T. Su, H.K. Lee, Anal. Chem. 77 (2005) 1988. [25] M.A. Jeannot, F.F. M.A. Anal. Chem. 69 (1997) 235. [26] M.C. Lopez-Blanco, S. Blanco-Cid, B. Cancho-Grande, J. Simal-Gandara, J. Chromatogr. A 984 (2003) 245. [27] P.S. Chen, S.D. Huang, J. Chromatogr. A 1118 (2006) 161. [28] M. Arias-Estevez, et al., Agriculture, Ecosystems and Environment 123 (2008) 247. [29] R.R. Otero, B.C. Grande, J.S. Gandara, J. Chromatogr. A 992 (2003) 121. [30] R.R. Otero, C.Y. Ruiz, B.C. Grande, J.S. Gandara, J. Chromatogr. A 942 (2002) 41. [31] M.C. Lopez-Blanco, B.C. Grande, J.S. Gandara, J. Chromatogr. A 963 (2002) 117. [32] M.C. Lopez-Blanco, B.R. Rodriguez, B.C. Grande, J.S. Gandara, J. Chromatogr. A 976 (2002) 293. [33] A.S. Yazdi, H. Assadi, Chromatographia 60 (2004) 699.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Dynamic headspace time-extended helix liquid-phase microextraction Shih-Pin Huang, Pai-Shan Chen, Shang-Da Huang ∗ Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan
a r t i c l e
i n f o
Article history: Received 11 November 2008 Received in revised form 9 February 2009 Accepted 13 March 2009 Available online 18 March 2009 Keywords: Headspace LPME Dynamic hollow fiber-LPME Organochlorine Pesticides GC–MS
a b s t r a c t Liquid-phase microextraction (LPME) has been proved to be a fast, inexpensive and effective sample pre-treatment technique for the analyses of pesticides and many other compounds. In this investigation, a new headspace microextraction technique, dynamic headspace time-extended helix liquid-phase microextraction (DHS-TEH-LPME), is presented. In this work, use of a solvent cooling system, permits the temperature of the extraction solvent to be lowered. Lowering the temperature of the extraction solvent not only reduces solvent loss but also extends the feasible extraction time, thereby improving extraction efficiency. Use of a larger volume of the solvent not only extends the feasible extraction time but also, after extraction, leaves a larger volume to be directly injected into the gas chromatography (GC) to increase extraction efficiency and instrument signal. The DHS-TEH-LPME technique was used to extract six organochlorine pesticides (OCPs) from 110 ml water samples that had been spiked with the analytes at ng/l levels, and stirred for 60 min. The proposed method attained enrichments up to 2121 fold. The effects of extraction solvent identity, sample agitation, extraction time, extraction temperature, and salt concentration on extraction performance were also investigated. The method detection limits (MDLs) varied from 0.2 to 25 ng/l. The calibration curves were linear for at least 2 orders of magnitude with R2 0.996. Relative recoveries in river water were more than 86%. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Organochlorine pesticides (OCPs) were introduced after World War II for the control of vector-borne diseases and were also used as insecticides in agriculture [1]. However, OCPs are known to accumulate in the tissues of living organisms, and their general worldwide extensive usage has led to substantial bioaccumulation. These compounds pose potential hazards to human health and the environment. OCPs have now been banned in most developed countries. However, the early spectacular success of dichlorodiphenyltrichloroethane (DDT) for malaria eradication in some countries has resulted in this compound’s continued use in developing countries. Studies have suggested that DDT and related compounds may affect the normal function of the endocrine system [2,28]. The ability of the prevalent isomer of the major and most persistent DDT derivative, p,p -dichlorodiphenyldichloroethylene (p,p -DDE), to bind to the androgen receptor in male rats has been reported [3,4]. For environmental and drinking water, the European Union (EU) has established the maximum admissible concentration of a single compound as 0.1 g/l, and 0.5 g/l is the maximum allowed for the total concentration of all OCPs [5,6]. These very low
∗ Corresponding author. Tel.: +886 3 572 1194; fax: +886 3 573 6979. E-mail address: [email protected] (S.-D. Huang). 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.03.030
concentrations make the development of an efficient and sensitive analytical method necessary. The detection of such contaminants is an important task for chemists working in the field of environment protection. Analyte extraction and pre-treatment is the most challenging and time-consuming step in an analytical procedure. Many extraction procedures are available viz. liquid–liquid extraction (LLE) and solid-phase extraction (SPE) techniques [29]. However, LLE is time consuming, generally labor intensive, and requires large quantities of expensive, toxic and environmentally unfriendly organic solvents. The conventional offline SPE technique requires less solvent, but is relatively expensive. In analytical chemistry, the trend is toward simplification and miniaturization of the sample preparation steps, and toward a decrease in the quantities of organic solvents used. More recently, a novel technique, solid-phase microextraction (SPME) [30–32], solvent microextraction (SME) [7,8] or single-drop microextraction (SDME)/liquid-phase microextraction (LPME) [9,10], has been developed. In 2002, Shen and Lee developed a LPME method based on a protected hollow-fiber membrane combined with the microsyringe and termed it hollow-fiber liquid-phase microextraction (HF-LPME) [11,12]. LPME has been shown to be an advantageous alternative sample preparation method to conventional LLE, and is relatively low in cost compared with SPE and SPME. LPME requires the use of quite small volumes (e.g., 5 l or less) of organic solvent to extract analytes from aqueous matrices.
4348
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Recently, two major variants have been developed for hollowfiber-protected LPME: direct-immersion and headspace. The latter, an analytical method for volatile organic compounds (VOCs) in aqueous or other matrices, is utilized widely as a method of directly detecting VOCs without the interference of a dirty matrix that could occur in the direct-immersion mode [13,14]. In such analysis, a more efficient sampling method is required if one is dealing with semivolatile compounds or with VOCs present at trace levels. However, the main problem of the technique is that all the commonly used organic solvents for GC analysis have high vapor pressures, which causes them to evaporate too rapidly in air. Moreover, when using water-miscible solvents in single drop microextraction, the increase in drop size that occurs during sampling may cause the drop to fall from the needle [15]. Apparently, the selection of suitable organic solvents is limited, as the vapor pressure of the solvent must be considered in addition to its extraction efficiency. Selection of a solvent which is compatible with GC and suitable for analysis in the headspace is a key step. Theis et al. [16] reported headspace single drop microextraction in which 1-octanol, a solvent with a very low vapor pressure, was used. Shen and Lee [17] have introduced dynamic headspace LPME to overcome some limitations of static headspace single drop microextraction. Jiang et al. [18] developed a dynamic hollow-fiber-supported headspace liquid-phase microextraction (DHF-HS-LPME) approach. A hollow fiber filled with the extracting solvent is placed onto a microsyringe needle and placed in the headspace of the sample solution. The solvent within the fiber is moved to-and-fro by means of a programmable syringe pump. However, there are some problems with these techniques. In DHF-HS-LPME, the extracting solvent within the straight fiber (1.5 cm) will stack and gradually form a droplet at the end of the fiber during extraction. This causes air bubbles to enter the fiber. Because of this phenomenon, it is difficult for the syringe pump to withdraw extracting solvent or to just withdraw discontinuous portions of solvent into a microsyringe after the extraction procedure. [19] Therefore, longer fibers which possess more extraction contact interface and solvent volume may not be suitable for use in headspace analysis. On the other hand, for headspace LPME, heating specimen solution apparently enhances the extraction efficiency, but it also hastens the evaporation of the limited amount of extracting solvent. Also, the solvent volume of a single drop microextraction or a DHF-HS-LPME cannot be too large. To address the above-mentioned problems, an efficient headspace LPME technique, dynamic headspace time-extended helix liquid-phase microextraction (DHS-TEH-LPME), is developed. Here, a long fiber (4.5 cm) is bent as a helix shape (Fig. 1). One end of the fiber is affixed to a pin and the other end is inserted into a microsyringe which contains 13 l of extracting solvent. The solvent is moved back and forth in the fiber by a programmable syringe pump. The exceed solvent that is held outside of the fiber forms a large droplet that is suspended at the ring portion of a helix-shaped fiber. As for the shape of the helix-type fiber, the extracting solvent does not stack at the end of the fiber as in the case of DHS-TEHLPME. Without interference of air bubbles, the solvent maintaining an unbroken form is withdrawn into the microsyringe easily and completely. Also, the long fiber and the large droplet provide a larger contact surface and larger volume of organic solvent. In DHS-TEHLPME analysis, the extraction time can be extended in the following two ways. In one of them, the solvent cooling assisting system, lowering the temperature of the extracting solvent reduces solvent loss. The other way is by increasing the initial volume of extracting solvent. Because solvent loss is reduced, this factor may be ignored during the dynamic headspace microextraction. Therefore, after extraction, a larger volume is available to be directly injected into the gas chromatography (GC). This improves extraction efficiency and signal amplitude.
Fig. 1. Experimental setup of SC-DHS-TEH-LPME.
Analytes are extracted into a microliter volume of extracting solvent, so analytes may potentially be enriched by factors of hundreds or thousands, depending on their distribution constants and the sample volume [20]. In this work, the large sample volumes (110 ml) used were intended to simulate ultra-trace level determinations of OCPs in river waters. The proposed method could attain enrichments of up to 2121-fold. In order to achieve the best extraction efficiency, the influence of parameters such as organic solvent, temperature, stirring rate, extraction time and ionic strength was investigated. The optimized DHS-TEH-LPME technique was then successfully applied to determine OCPs in water samples. Detection was performed by GC–MS.
2. Experimental 2.1. Reagents and materials All chemicals were of reagent grade and were used without further purification. Heptachlor, aldrin, endosulfan, p,p -DDE, dieldrin, and o,p-DDT were purchased from Fluka (Buchs, Switzerland). Stock standard solutions of the studied pesticides were prepared in methanol at a concentration of 0.1 mg/ml and stored at 4 ◦ C. Working solutions used to optimize the parameters in LPME were prepared at a concentration of 10 ng/ml every day. Purified water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). Sodium chloride was bought from Showa Chemicals (Tokyo, Japan). The extracting solvents 1-octanol, 1hexanol, 1-pentanol, nonane and o-xylene were purchased from Fluka (analytical-grade). A Q 3/2 Accurel Polypropylene hollow-fiber membrane (600 m I.D., 200 m wall thickness, 0.2 m pore size) was purchased from Membrana GmbH (Wuppertal, Germany). The hollow fiber was cut into 4.5 cm segments and was cleaned in acetone and dried before use.
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
4349
2.2. Instrumentation Analyses were performed using a Varian (Walnut Creek, CA, USA) CP-3800 gas chromatography system with a 1079 injector kept at 260 ◦ C, in splitless mode (1 min). A ChromatoProbe (Varian Instruments, Walnut Creek, CA, USA) was coupled with this injector. A 30m DB-5MS (0.25 mm I.D., 0.25 mm film thickness) fused-silica capillary column (J&W Scientific, Folsom, CA, USA) was used for all analyses; helium at a constant flow of 1 ml/min was the carrier gas. The temperature was held at 150 ◦ C for 2 min, increased to 230 ◦ C at 20 ◦ C/min and held for 2 min, then raised to 280 ◦ C at 20 ◦ C/min and held for 6 min. A Varian Saturn 2000 ion trap GC–MS system was used for tandem mass spectrometric (MS–MS) detection. The ion trap mass spectrometer was operated in the electron ionization (EI) mode; the ionization energy was set at 70 eV and the source temperature at 200 ◦ C. Full-scan MS–MS data were acquired over the m/z range 100–450 to obtain the fragmentation spectra of the analytes. The Varian-supplied routine, ‘Automated Methods Development (AMD)’, was used to find the optimum collision-induced dissociation (CID) excitation amplitude. Specific product ions in the MS–MS spectra were used for detection in this study, and the appropriate extracted ion chromatograms were calculated and integrated for quantitation. The qualitative GC–MS–MS data for the pesticides studied are shown in Table 1. 2.3. Extraction process A 3 cm × 0.6 cm stir bar was put into 110 ml of sample solution in a 120 ml home made vial. The syringe was mounted in a cooling assisting system and was fitted on the syringe pump. A headspace sample vial septum cap was pierced by the syringe. A 13 l portion of 1-octanol was drawn into the microsyringe. The needle tip was inserted into the 4.5 cm hollow fiber. The hollow fiber was revolved around the syringe needle, which formed in a helix shape. The end of the fiber was fixed on a pin, as shown in Fig. 1. Subsequently the assembly was then immersed in 1-octanol to impregnate the pores of the hollow fiber for 15 s to impregnate the pores of the fiber. Then the prepared fiber was removed and put in the headspace above a 110 ml aqueous sample which contained 10 g/l of the target compounds. The syringe pump was pre-programmed to dispose and withdraw 10 l of the organic phase in a continuous mode at a speed of 40 l/min and the cooling system was set at −10 ◦ C. The solution was then stirred at 600 rpm at 50 ◦ C for 30 min. The final movement of the plunger was to withdraw the extract into the syringe barrel and injected 10 l into the GC–MS for analysis. The used fiber was discarded and a fresh one was used for the next experiment. 3. Results and discussion 3.1. Selection of organic solvent Selection of a suitable organic extraction solvent is crucial for dynamic headspace LPME. The selection criteria for an organic solvent to be a suitable choice are that (1) it should be easily immo-
Fig. 2. Effect of extraction solvent on LPME efficiency (n = 3), spiked to 10 g/l of each of the analytes, 600 rpm, 50 ◦ C, 30 min.
bilized in the pores of the polypropylene hollow fiber, and (2) the organic solvent should be of low volatility to prevent solvent loss, (3) volatile enough for use in GC–MS analysis. In consideration of these factors, o-xylene, nonane, 1-pentanol, 1-hexanol and 1octanol were tested for their suitability. Fig. 2 exhibits the extraction efficiencies for the tested organic solvents. According to Fig. 2, the extraction efficiency was best when 1-octanol was used for most OCPs except p,p -DDE. The reason could be that the 1-octanol, 1hexanol and 1-pentanol are more polar than o-xylene and nonane and high polarity enhances better dissolution because OCPs are polar compounds. However, the volatility of 1-octanol is lower than that of 1-hexanol and 1-pentanol. The lower volatility the solvent has, the less volume the solvent will lose the longer the extraction time will take in the process. Therefore the selection of organic solvent is more flexible for this approach than is the case for the previous headspace LPME method. Hence for further experiments 1-octanol was chosen as the organic extractant. 3.2. Effect of stirring speed HS-LPME is based on equilibrium partitioning of the analyte between three phases: the condensed aqueous phase, the headspace phase and the organic solvent phase. Mass transfer in the headspace is assumed to be a faster process than it is in either of the condensed phases since diffusion coefficients in the gas phase are typically about 104 times greater than corresponding diffusion coefficients in the condensed phase. Thus, both slow mass transfers in the aqueous phase and diffusion of analyte into the organic solvent phase are limiting steps in the overall extraction process [21]. The aqueous phase mass transfer coefficient of solute increases with increasing stirring rate. Therefore, agitation of the sample solution improves mass transfer in the aqueous phase and induces convection in the headspace, so consequently the equilibrium between the aqueous and vapor phase can be attained more rapidly. Fig. 3 describes the influence of stirring speed on the extraction efficiency. The extraction efficiency increased with increasing stirring rate up
Table 1 Qualitative GC/MS/MS data for the studied pesticides. Analyte
Retention time (min)
Precursor ion (m/z)a
CID excitation amplitude (V)
Product ions, m/z (Relative abundance (%))
Quantitation ions (m/z)
Heptachlor Aldrin Endosulfan p,p -DDE Dieldrin o,p -DDT
6.04 6.57 7.65 7.85 8.02 8.49
272 293 241 318 279 235
0.32 0.33 0.40 0.38 0.40 0.47
272(12), 270(100), 237(79) 293(25), 291(100), 257(56) 241(27), 239(100), 206(45) 318(37), 316(100), 283(17) 279(28), 277(100), 243(34) 235(21), 200(100), 165(39)
270/237 291/257 239/206 316/283 277/243 200/165
a
Not the molecular ion, selected on basis of signal intensity.
4350
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Fig. 3. Effect of stir rate on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 50 ◦ C, 30 min.
Fig. 5. Effect of temperature on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 60 min, stirring at 800 rpm.
to 800 rpm. However, when the stirring rate exceeded 800 rpm, the stirring bar in the sample solution could not move steadily, and consequently lowered the extraction efficiency. Finally, a stirring speed of 800 rpm was used for subsequent experiments.
sample matrix/headspace/hollow-fiber-supported organic phase. The effect of extraction temperature was evaluated from 50 to 90 ◦ C. As seen in Fig. 5, the extraction efficiency of OCPs increases with increasing extraction temperature. When the sample solution was heated to a temperature higher than 80 ◦ C, extraction efficiency dropped, which may due to the loss of extracting solvent, and the competition between Henry’s constant and the partition coefficient. At higher temperatures, both the Henry’s constants and the diffusion coefficients of the target analytes become larger. This results in an increase of the vapor pressure of the OCPs. Therefore their concentrations also rise in the headspace [23]. However, the partition coefficient between the headspace and the organic phase is reduced with increasing temperature [24]. For these reasons, a temperature of 80 ◦ C was used for further work.
3.3. Effect of extraction time Dynamic headspace LPME is not an exhaustive extraction technique; rather, it is an equilibrium processes in which analytes partition between the condensed phase/headspace, and the headspace/extracting organic solvent. Hence, extraction efficiency depends on the length of the extraction time. Fig. 4 shows the influence of extraction time on efficiency of the OCPs. The peak areas increase with increasing extraction time in the range of 30–65 min. However, longer extraction times (>65 min) may lead to too much water vapor condensation which may freeze in the syringe needle and interfere with efficient extraction. On the other hand, exposure time in the extraction step should not be excessively long, as would be the case if one were trying to reach equilibrium; extraction time should be just long enough to give satisfactory sensitivity and precision [9,22]. Thus the extraction time was set at 60 min in the following experiments.
3.5. Effect of ionic strength
For headspace analysis of semi-volatile compounds, another important parameter is temperature. Temperature has a significant effect on both the kinetics and the thermodynamics of the extraction process. Temperature affects the kinetics of sorption in the extracting organic solvent by determining the vapor pressures of analytes and diffusion coefficient values in the three phases [21]
The effect of ionic strength on extraction efficiency was studied by preparing standard solutions of the analytes together with NaCl at concentrations varying from 0 to 30% (w/v). Adding salt to the aqueous samples may have several effects on the analysis. In generally, a salting out effect has been observed in LLE and solid-phase microextraction (SPME). Salt can decrease the solubilities of analytes in aqueous samples and can enhance their partitioning into the organic phase (LLE) or the adsorbent (SPME). However, the presence of salt was observed to limit the extraction of analytes by some authors [18,25–27]. In our study, as shown in Fig. 6, no increase in extraction was observed after the addition of the sodium chloride. On the contrary, the extraction efficiencies were highest without addition of the sodium chloride, and subsequently decreased as more was added. This is due to the fact that polar molecules may
Fig. 4. Effect of extraction time on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 50 ◦ C.
Fig. 6. Effect of NaCl concentration on LPME efficiency (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 60 min, 800 rpm, 80 ◦ C.
3.4. Effect of temperature
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Fig. 7. Comparisons of SC-DHS-TEH-LPME performance with DHS-TEH-LPME (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 80 ◦ C. The extraction time of DHS-TEH-LPME was set at 30 min, and SC-DHS-TEHLPME was set at 30 and 60 min.
participate in electrostatic interactions with the salt ions in solution [33], thereby reducing the rate of mass transfer. Hence, further extractions were made without the addition of NaCl. 3.6. Effect of solvent cooling assisting system The solvent cooling assisting system was used to lower the temperature of the organic solvent on the fiber, which prevents extracting solvent from being lost by evaporation at higher temperatures. At too high a temperature (>0 ◦ C), the effects of the cooling system could not be readily observed. At too low a temperature (<−20 ◦ C), the viscosity of the solvent increases, apparently, which might lead to a smaller amount of extraction volume withdrawn by the syringe. Based on these reasons a cooling temperature of −10 ◦ C was used for further studies.
4351
Fig. 8. Comparisons of DHS-TEH-LPME performance with SC-DHS-TEH-LPME (n = 3), 1-octanol as extraction solvent, spiked to 10 g/l of each of the analytes, 800 rpm, 80 ◦ C. The extraction time of SC-DHS-TEH-LPME was set at 30 min, and DHS-TEHLPME was set at 30–60 min.
With the solvent cooling assisting system, the temperature of the extraction solvent could be lowered, which lowers the vapor pressure, decreases the loss of solvent, and permits one to extend the extraction time. Comparison between extractions with and without the solvent cooling assisting system is shown in Fig. 7. Without the solvent cooling assisting system, if we want to maintain the GC input amount at 10 l, the longest extraction possible is 30 min. However, with the solvent cooling assisting system, the extraction time can be extended to 60 min. On the other hand, for a given extraction time, it was observed that, as expected, at a lower solvent temperature the extraction yield was increased. The main reason is that the process of analyte absorption in the solvent is exothermic and the partition coefficient is temperature dependent [21]. If a low solvent temperature is maintained during sampling, the partition coefficient is increased. As a result, greater mass transfer of analyte into the solvent occurs and sensitivity significantly increases.
Fig. 9. Chromatogram of OCPs obtained by LPME under optimized conditions (A) unspiked river water; (B) spiked 1.0 g/l in the river water. Peaks: (1) heptachlor; (2) adrin; (3) endosulfan; (4) p,p -DDE; (5) dieldrin; (6) o,p -DDT.
4352
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353
Table 2 Performance of the validation analysis. Pesticides Heptachlor Aldrin ␣-Endosulfan p,p -DDE Dieldrin o,p -DDT
Linearity (g/l) 0.01–1 0.01–1 0.1–50 0.01–1 0.1–50 0.01–1
R2
MDLsa (ng/l)
RSD (%)b (n = 5)
Absolute Recoveryc (%)
Relative Recoveryd (%)
0.999 0.998 0.997 0.998 0.996 0.997
0.42 0.33 19 0.27 25 0.37
16.3 15.4 12.5 17.8 16.3 17.4
6.2 10.1 5.8 19.3 10.8 13.8
86 106 109 104 106 98
± ± ± ± ± ±
10 17 5 15 12 15
Enrichment Factorb (fold) 677 1112 633 2121 1184 1520
a MDL are calculated as three times the standard deviation of seven replicated runs of standard solution: amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. b Amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. c A percentage value obtained considering spiked amount in the river water as 100%; amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. d A percentage value obtained considering extraction yields in the river water as 100%; amount of analytes added was 1.0 g/l for each analyte; mean ± standard deviation (n = 3).
Table 3 Comparisons of performance of DHS-TEH-LPME and SC-HF-LPME-GC–MS in this study. Pesticides
Heptachlor Aldrin ␣-Endosulfan p,p -DDE Dieldrin o,p -DDT a b
DHS-TEH-LPME
SC-HF-LPME-GC–MS
MDLs (ng/l)
RSDa (%)
Enrichment Factor (fold)
MDLs (ng/l)
RSDb (%)
Enrichment Factor (fold)
0.42 0.33 19 0.27 25 0.37
16.3 15.4 12.5 17.8 16.3 17.4
677 1112 633 2121 1184 1520
42 51 125 53 209 44
14.2 5.5 15.2 9.7 14.1 12.3
141 125 211 65 100 82
Amount of analytes added: Heptachlor, 1 ng/l; Aldrin, 1 ng/l; ␣-Endosulfan, 50 ng/l; p,p -DDE, 1 ng/l; Dieldrin, 50 ng/l; o,p -DDT, 1 ng/l. Amount of analytes added was 1.0 g/l for each analyte.
3.7. Comparisons of DHS-TEH-LPME performance with SC-DHF-HS-LPME SC-DHF-HS-LPME was evaluated with the same extraction conditions as DHS-TEH-LPME (expect extraction time) for water samples spiked at a pesticide concentration of 10 g/l; the results are shown in Fig. 8. It could be directly observed from the experiment of SC-DHF-HS-LPME that when the extraction time took more than 30 min, there would be water vapor condensed in the syringe needle. It may lead to the extraction efficiency to decline. Thus the extraction time of SC-DHF-HS-LPME was set at 30 min. At the same extraction time, the peak area of DHS-TEH-LPME was higher than SC-DHF-HS-LPME. It was highly possible reason that DHS-TEHLPME provides a larger contact surface of organic solvent. On the other hand, the extraction time of DHS-TEH-LPME could last longer, so its extraction efficiency is better. 3.8. Quantitative aspects The developed procedure was applied to the determination of OCPs in the river water. A general procedure was followed to extract the target analytes. Fig. 9 shows typical chromatograms obtained from the analysis of the river water. No OCPs was found in the unspiked river water. Repeatability, linearity, correlation coefficient, detection limit and enrichment factors were investigated under optimized experimental conditions. The performance of the developed procedure is summarized in Table 2. It can be seen that the RSD values are smaller than 17.8% based on the peak areas for five replicate runs. This large value is probably due to the fact that large sample volumes (110 ml) were used. Each analyte exhibited a good correlation coefficient, with R2 ≥ 0.996. Relative recoveries, determined as the ratio of the concentration found in the river water and in deionized water samples, spiked with the same amount of analytes were calculated. The relative recoveries of spiked OCPs at 1.0 g/l were between 86 and 109%. The method detection limits (MDLs) values were calculated as three times the standard deviation
of seven replicate runs of river water spiked with low concentrations of the analytes. The MDLs range from 0.27 to 0.42 ng/l, except ␣-endosulfan (19 ng/l) and dieldrin (25 ng/l). The enrichment factors, defined as the ratio between the final (equilibrium) concentration of analyte in the organic phase and the initial concentration of analyte in the aqueous sample, were between 633 and 2121. In comparison with a previously reported headspace liquidphase microextraction method [19] in Table 3, we obtained much lower MDLs and higher enrichment for all targeted analytes. 4. Conclusions In the present study a high-enrichment sample preparation method for the analysis of trace organochlorine pesticides, based on a dynamic headspace time-extended helix liquid-phase microextraction technique has been developed. This method involves the use of only ∼13 l of organic solvent for each extraction. The potential for a dynamic liquid-phase microextraction technique using a commercially available syringe pump has been demonstrated by analyzing OCPs from river water. The method provides extracts with high-enrichment factors along with excellent sample clean-up. Thus the method provides preconcentration and sample clean-up simultaneously. However, the greatest weakness of DHS-THE-LPME was that it could not be fully automated. Besides, due to the greater volume of the sample solution, the time for extraction equilibrium took longer. Thus, it led to higher RSD. Acknowledgment The support of the National Science Council, Taiwan (NSC 962113-M-007-030-MY3) is acknowledged. References [1] H.P. Lia, G.C. Lib, J.F. Jen, J. Chromatogr. A 1012 (2003) 129. [2] E. Hileman, Chem. Eng. News 31 (January) (1994) 19.
S.-P. Huang et al. / J. Chromatogr. A 1216 (2009) 4347–4353 [3] W.R. Keice, R.C. Stone, S.C. Law, L.E. Gray, J.A. Kemppainen, E.M. Wilson, Nature 375 (1995) 581. [4] G.R. van der Hoff, P. van Zoonen, J. Chromatogr. A 843 (1999) 301. [5] European Union, Council Directive 98/83/EC on the quality of water intended for human consumption, Off. J. Eur. Commun. 32 (1998) 330. [6] National Primary Drinking Water Regulation (40 CFR Part 141), Fed. Regis. 19 (2002) 426. [7] M.A. Jeannot, F.F. Cantwell, Anal. Chem. 68 (1996) 2236. [8] M. Ma, F.F. Cantwell, Anal. Chem. 70 (1998) 3912. [9] Y. He, H.K. Lee, Anal. Chem. 69 (1997) 4634. [10] Y. Wang, Y.C. Kwok, Y. He, H.K. Lee, Anal. Chem. 70 (1998) 4610. [11] G. Shen, H.K. Lee, Anal. Chem. 74 (2002) 648. [12] L. Zhao, H.K. Lee, Anal. Chem. 74 (2002) 2486. [13] J. Drozd, J. Novak, J. Chromatogr. 165 (1979) 141. [14] A.J. Nunez, L.F. Gonzalez, J. Janak, J. Chromatogr. 300 (1984) 127. [15] A. Tankeviciute, R. Kazlauskas, V. Vickackaite, Analyst 126 (2001) 1674. [16] A.L. Theis, A.J. Waldack, S.M. Hansen, M.A. Jeannot, Anal. Chem. 73 (2001) 5651. [17] G. Shen, H.K. Lee, Anal. Chem. 75 (2003) 98. [18] X.M. Jiang, C. Basheer, H. Zhang, H.K. Lee, J. Chromatogr. A 1087 (2005) 289.
4353
[19] S.P. Huang, S.D. Huang, J. Chromatogr. A 1176 (2007) 19. [20] S.H. Tung, V. Terje, A. Trude, J. Einar, E.R. Knut, P.-B. Stig, Anal. Chim. Acta 592 (2007) 1. [21] A. Przyjazny, J.M. Kokosa, J. Chromatogr. A 977 (2002) 143. [22] M.H. Ma, F.F. Cantwell, Anal. Chem. 71 (1999) 388. [23] M. Saraji, J. Chromatogr. A 1062 (2005) 15. [24] J. Zhang, T. Su, H.K. Lee, Anal. Chem. 77 (2005) 1988. [25] M.A. Jeannot, F.F. M.A. Anal. Chem. 69 (1997) 235. [26] M.C. Lopez-Blanco, S. Blanco-Cid, B. Cancho-Grande, J. Simal-Gandara, J. Chromatogr. A 984 (2003) 245. [27] P.S. Chen, S.D. Huang, J. Chromatogr. A 1118 (2006) 161. [28] M. Arias-Estevez, et al., Agriculture, Ecosystems and Environment 123 (2008) 247. [29] R.R. Otero, B.C. Grande, J.S. Gandara, J. Chromatogr. A 992 (2003) 121. [30] R.R. Otero, C.Y. Ruiz, B.C. Grande, J.S. Gandara, J. Chromatogr. A 942 (2002) 41. [31] M.C. Lopez-Blanco, B.C. Grande, J.S. Gandara, J. Chromatogr. A 963 (2002) 117. [32] M.C. Lopez-Blanco, B.R. Rodriguez, B.C. Grande, J.S. Gandara, J. Chromatogr. A 976 (2002) 293. [33] A.S. Yazdi, H. Assadi, Chromatographia 60 (2004) 699.