Anal. Chem. 2000, 72, 4462-4467
Continuous-Flow Microextraction Exceeding1000-Fold Concentration of Dilute Analytes Wuping Liu and Hian Kee Lee*
Department of Chemistry, National University of Singapore, 3 Science Drive 3, Republic of Singapore 117543
A novel liquid-liquid microextraction method, that we have termed continuous-flow microextraction (CFME), is described. In a 0.5-mL glass chamber, an organic drop (1-5 µL) is held at the outlet tip of a polyetheretherketone (PEEK) connecting tubing which is immersed in a continuously flowing sample solution and acts as the fluid delivery duct and as a solvent holder. Extraction takes place between the organic drop and the flowing sample solution that is continuously ejected out of the PEEK tubing. Concentration factors of between 260- to 1600fold are achieved within 10 min of extraction. Aspects relevant to CFME were studied. In combination with gas chromatography-electron capture detection, CFME allows analytes to be detected at femtogram-per-milliliter levels. The performance of this technique was evaluated on the basis of the analysis of trace nitroaromatic compounds and chlorobenzenes in environmental samples. A variety of means are applicable to increase the performance of analyte extraction procedures. In conventional liquid-liquid extraction, it is common to agitate the mixture consisting of aqueous sample and extraction agent (usually a water-immiscible organic solvent) by magnetic stirring or by mechanical shaking.1 Because of the external forces imposed, the extraction agent is broken up into tiny liquid drops, leading to an increase in its contact area with the aqueous sample; also, mass transfer is accelerated in the bulk aqueous phase, the organic phase, and the interfacial film. Consequently, extraction is speeded up. However, being limited by the phase ratio V0/Vaq (where V0 and Vaq are volumes of extraction agent and aqueous solution, respectively), concentration factors achieved by conventional liquid-liquid extraction are usually poorer than those obtained by other techniques, such as solid-phase extraction (SPE),2 solidphase microextraction (SPME),3 supported liquid membrane (SLM) extraction,4 field-amplified stacking injection (FASI),5,6 or analyte sweeping in micellar electrokinetic chromatography.7 * Corresponding author. E-mail:
[email protected] Fax: (65)-779-1691. (1) Thornton, J. D. Science and Practice of Liquid-Liquid Extraction. Vol. 1; Oxford University Press: Oxford, UK, 1992; Chapter 5. (2) Liu, W. P.; Lee, H. K. Talanta 1998, 45, 631-639. (3) Zhang, Z. Y.; Yang, M. J.; Pawliszyn, J. Anal. Chem. 1994, 66, 844A853A. (4) Pa´lmarsdo´ttir, S.; Thordarson, E.; Edholm, L. E.; Jo¨nsson, J. Å.; Mathiasson, L. Anal. Chem. 1997, 69, 1732-1737. (5) Chien, R. L.; Burgi, D. S. Anal. Chem. 1992, 64, 489A-496A. (6) Liu, W. P.; Lee, H. K. Anal. Chem. 1998, 70, 2666-2675.
4462 Analytical Chemistry, Vol. 72, No. 18, September 15, 2000
Recently, solvent microextraction, or liquid-phase microextraction (LPME), has attracted increasing attention. A microextraction method was described in which a small drop (8 µL) of a water-immiscible organic solvent was held at the end of a Teflon rod that was immersed in a stirred aqueous sample solution.8 Improvement was made to the original method by using a microsyringe, in place of the Teflon rod, as the solvent holder.9 Microextraction was carried out into a 1-µL solvent drop at the tip of the syringe needle.9,10 It is noted that, in solvent microextraction, the solvent drop cannot be further broken up; stirring is mainly to accelerate the kinetics of extraction by minimizing the thickness of the interfacial film. Alternative extraction modes were described in dynamic LPME.11,12 Extraction was performed between microliters of aqueous sample and microliters of extraction agent, by repetitively pulling and pushing the plunger within the glass barrel of a microsyringe. Inherent from the use of solvent drops at microlevels, and thus relatively low phase ratios, a high concentration factor is expected in LPME. To date, the potential of microextraction techniques has yet been fully exploited, in both methodology and application.9,11 In this study, we report a novel liquid-liquid microextraction technique that we term continuous-flow microextraction (CFME). A polyetheretherketone (PEEK) tubing, now commonly used in HPLC plumbing, is used as a “holder” of the extraction solvent drop (Figure 1). The extraction solvent (∼3 µL) is introduced through a conventional HPLC injection valve. It moves with the pumped sample solution and reaches the outlet end of the PEEK tubing. Behaving differently from the bulk sample solution, which flows right through the tubing and the extraction unit (made of glass) to the waste, the extraction solvent remains near the tip of the PEEK tubing outlet as a solvent drop. The virtually immobilized solvent drop permits continuous interaction with the flowing sample solution and, simultaneously, extraction proceeds. Aspects relevant to the continuous-flow microextraction process were studied. Preconcentration performance and potential applications of the present technique were evaluated. (7) (8) (9) (10) (11) (12)
Quirino, J. P.; Terabe, S. Science (Washington, D.C.) 1998, 282, 465-468. Jeannot, M. A.; Cantwell, F. F. Anal. Chem. 1996, 68, 2236-2240. Jeannot, M. A.; Cantwell, F. F. Anal. Chem. 1997, 69, 235-239. Jeannot, M. A.; Cantwell, F. F. Anal. Chem. 1997, 69, 2935-2940. He, Y.; Lee, H. K. Anal. Chem. 1997, 69, 4634-4640. Wang, Y.; Kwok, Y. C.; He, Y.; Lee, H. K. Anal. Chem. 1998, 70, 46104614. 10.1021/ac000240x CCC: $19.00
© 2000 American Chemical Society Published on Web 08/10/2000
Figure 1. Assembly of continuous-flow microextraction system. (1) Connecting PEEK tubing, inserted into the extraction chamber; (2) Modified pipet tip; (3) “o”-ring; (4) Inlet of extraction chamber; (5) Extraction chamber; (6) microsyringe; (7) solvent drop. Table 1. Performance of Continuous-Flow Microextraction analyte (abbreviation) 2-nitrotoluene (2-NT) 4-nitrotoluene (4-NT) 1,4-dinitrobenzene (1,4-DiNB) 1,3-dinitrobenzene (1,3-DiNB) 1,2-dinitrobenzene (1,2-DiNB) 1,2-dichlorobenzene (1,2-DiCB) 1,3-dichlorobenzene (1,3-DiCB) 1,3,5-trichlorobenzene (1,3,5-TriCB) 1,2,3-trichlorobenzene (1,2,3-TriCB) 1,2,4,5-tetrachlorobenzene (1,2,4,5-TetCB) pentachlorobenzene (PentCB)
concn linear factor range % % LOD (-fold) (pg/mL) RSDa recovery (pg/mL) 510 400 790
10-1000 20-1000 10-1000
8.8 7.6 6.4
93.2 93.1 97.1
5 10 5
380
10-1000
6.9
91.2
10
590
10-1000
6.3
102.5
5
560
5-500
7.2
104.3
0.2
700
5-500
6.8
96
0.2
940
1-100
4.7
98.7
0.1
1400
1-100
4.6
99.1
0.1
260
1-100
5.1
91.5
0.5
1600
0.5-50
3.8
102.2
0.01
a Analyte concentrations: 0.1 ng/mL of 2-NT, 1,2-DiNB, 1,3-DiNB, and 1,4-DiNB; 0.2 ng/mL of 4-NT; 5 pg/mL of PentCB; 10 pg/mL of 1,2,3-TriCB, 1,3,5-TriCB, and 1,2,4,5-TetCB; and 50 pg/mL of 1,2-DiCB and 1,3-DiCB.
EXPERIMENTAL SECTION Chemicals and Samples. Five nitroaromatic compounds and six chlorinated benzenes (Table 1) were bought from Aldrich (Milwaukee, WI). Stock solutions (1 mg/mL) of analytes were prepared in methanol separately. HPLC-grade methanol, toluene, isooctane, chloroform, n-butyl acetate, xylene, and cyclohexane were purchased from J. T. Baker (Phillipsburg, NJ). Prior to use, the solvents were redistilled to remove traces of water. Deionized water was produced on a Milli-Q (Millipore, Bedford, MA) water purification system. Water samples were prepared by spiking deionized water with analytes at known concentrations (500 fg/mL to 500 pg/mL). To study extraction performance under different conditions, two working solutions (prepared in methanol) were used. One consisted of five nitroaromatic compounds (2-NT, 1,4-DiNB, 1,3DiNB, and 1,2-DiNB, 1.0 ng/mL for each; and 4-NT, 2.0 ng/mL) (see Table 1 for explanations of abbreviations). The other consisted of 6 chlorobenzenes (1,2-DiCB, 50 pg/mL; 1,3-DiCB,
50 pg/mL; 1,2,3-TriCB, 10 pg/mL; 1,3,5-TriCB, 10 pg/mL; 1,2,4,5TetCB, 10 pg/mL; and PentCB, 5 pg/mL) (see Table 1 for explanations of abbreviations). Surface seawater samples (1.0 L each) were collected into clean glass bottles from the south coast of Singapore (50 m from the beach). The fresh samples were filtered through a no. 4 sintered funnel, followed by storage (24 h) at room temperature (25 °C) to allow settling of suspended matter. Filtration of the samples was conducted again prior to extractions. Apparatus. Extraction System. A JASCO (Tokyo, Japan) PU980 HPLC pump and a Rheodyne (Cotati, CA) 7125 injector (equipped with a 5-µL sample loop) were used for sample delivery and extraction solvent introduction, respectively. Extraction was performed in a homemade glass chamber (∼0.97. The recoveries (%rec, three replicates) and reproducibility (%RSD, five replicates) experiments were carried out on spiked water samples. The limits of detection (LOD) for the above analytes were calculated at a signal-to-noise ratio of 3. The linear range, %RSD, %rec, and LODs are given in Table 1. Detection limits at femtogram-per-milliliter levels can be achieved using CFME in combination with GC-ECD. However, RSD values >5% were observed for most of the analytes. This possibly resulted from the lack of an internal standard and the fact that peak areas used for calculation had not been corrected. It is possible improvements in reproducibility can be made by the automation of post-extraction solvent collection. Real-World Sample Analysis. The present technique was combined with GC-ECD to analyze trace environmental pollutants in surface seawater. Typical chromatograms generated by using toluene and isooctane as extraction solvents are presented in Figure 5 parts a and b, respectively. Six chlorobenzenes were identified in the samples. Their concentrations were as follows: 1,2-dichlorobenzene, 105 pg/mL; (1,3-dichlorobenzene), 51.2 pg/ mL; (1,2,3-trichlorobenzene), 2.26 pg/mL; (1,3,5-trichlorobenzene), 6.17 pg/mL; (1,2,4,5-tetrachlorobenzene), 4.12 pg/mL; and (pen-
tachlorobenzene), 1.82 pg/mL, respectively. The five nitroaromatic compounds were not found using this method, however. For comparison, the performance of dynamic-LPME,12 which has been demonstrated to give high preconcentration factors, was also evaluated (Figures 5c and 5d). Generally, better detection of chlorobenzenes, as well as higher total extraction performance, was achieved by using CFME in combination with GC-ECD than that obtained by dynamic LPME in combination with GC-ECD. Toluene was found to be a better solvent than isooctane in CFME for the extraction of chlorobenzenes, in comparison with isooctane being more favorable in dynamic-LPME of the same class of compounds. This is in agreement with data reported previously.12 CONCLUSIONS Continuous-flow microextraction is a single-step extraction method. It differs from other extraction approaches in the manner of extraction and preconcentration. In CFME, only a single drop of solvent fully and continuously makes contact with the sample solution. Both diffusion and molecular momentum that result from mechanical force (due to HPLC pumping) are expected to contribute to its effectiveness. With the use of an injector, precise control of solvent drop size can be achieved and introduction of undesirable air bubbles9,11 avoided. Extraction is affected by the type of tubing material, solvent type, flow rate of the sample solution, and the extraction time. Under optimum conditions, CFME permits preconcentration exceeding 1000 fold of dilute analytes. The combination of CFME with GC-ECD allows limits of detection at the femtogram-per-milliliter levels and provides a promising technique in the detection of trace amount of analytes with low water solubility, e.g., nitroaromatic compounds and chlorobenzenes in environmental samples. Another advantage is that because of the high preconcentration factor that can be achieved, smaller volumes of aqueous samples are needed for extraction (3 mL or less in the present work), in comparison with, for example, liquid-liquid extraction, in which hundreds of milliliters or even liters of samples are necessary. Notwithstanding the effectiveness of CFME, mechanisms relating to extraction and achievement of high extraction performance in CFME are still open for exploration. To date, numerous models have been designed to describe thermodynamics and kinetics of a liquid droplet at different states, such as the pendant state, the dispersed state, and the state that exists during a freefalling process.1 Mass transfer kinetics have also been studied by Cantwell and co-workers, on solvent microextraction into a single drop.8-10 CFME is distinguished from dynamic-LPME or solvent microextraction by the motion of bulk sample solution, and possibly, the extraction mechanism. Further investigations can focus on the theoretical aspect of CFME. Also, attention can be paid to expanding the applications of CFME to different aqueous sample matrixes and to other analyte classes such as pesticides, water-immiscible hormones, and steroids. Received for review February 28, 2000. Accepted June 14, 2000. AC000240X
Analytical Chemistry, Vol. 72, No. 18, September 15, 2000
4467
