Anal. Chem. 2004, 76, 855-858
Liquid-Phase Microextraction as an On-Line Preconcentration Method in Capillary Electrophoresis Kihwan Choi,† Yongseong Kim,‡ and Doo Soo Chung*,†
School of Chemistry, Seoul National University, Seoul 151-747, Korea, and Department of Chemistry, Kyungnam University, Masan 631-701, Korea
A simple and efficient sample preconcentration method for capillary electrophoresis has been developed using liquid-phase microextraction (LPME). A thin layer of an organic liquid was used to separate a drop of the aqueous acceptor phase hanging at the inlet of a capillary from the bulk aqueous donor phase. The donor-phase pH was 1.0, and the acceptor phase pH was 9.5. This pH difference caused the preconcentration of the acidic compounds, fluorescein and fluorescein isothiocyanate, into the acceptor-phase drop. Enrichment factors of 3 orders of magnitude were obtained with 30-min LPME at 35 °C. To overcome the poor concentration sensitivity of capillary electrophoresis (CE), on-column concentration methods such as sample stacking,1 field-amplified sample injection,2,3 large-volume stacking,4,5 transient isotachophoresis,6,7 and sweeping8,9 have been widely used. The advantage of on-column concentration is that no mechanical modification of the column is required and the concentration procedure is relatively simple with little manipulation of buffer and sample systems. However, its application is limited to cases satisfying specific requirements for the run buffer and sample solutions. As an alternative to these techniques, sample preconcentration methods coupled with CE can be employed. Solid-phase extraction, which was initially introduced to CE as a pretreatment technique, showed concentration effects, although a solvent desorption step was required and sample recovery was not reproducible.10,11 Liquid-liquid extraction (LLE) has been applied to GC,12,13 LC,14 and CE.15 Various liquid-phase microextraction (LPME) schemes achieving higher concentration * Corresponding author. Tel: +82-2-880-8130. Fax: +82-2-877-3025. E-mail:
[email protected]. † Seoul National University. ‡ Kyungnam University. (1) Burgi, D. S.; Chien, R. L. Anal. Chem. 1991, 63, 2042-2047. (2) Chien, R. L.; Burgi, D. S. J. Chromatogr. 1991, 559, 141-152. (3) Chien, R. L.; Burgi, D. S. J. Chromatogr. 1991, 559, 153-161. (4) He, Y.; Lee, H. K. Anal. Chem. 1999, 71, 995-1001. (5) Lee, J. H.; Choi, O. K.; Jung, H. S.; Kim, K. R.; Chung, D. S. Electrophoresis 2000, 21, 930-934. (6) Foret, F.; Szoko, E.; Karger, B. L. J. Chromatogr. 1992, 608, 3-12. (7) Foret, F.; Szoko, E.; Karger, B. L. Electrophoresis 1993, 14, 417-428. (8) Quirino, J. P.; Terabe, S. Science 1998, 282, 465-468. (9) Quirino, J. P.; Terabe, S. Anal. Chem. 1999, 71, 1638-1644. (10) Beattie, J. H.; Self, R.; Richards, M. P. Electrophoresis 1995, 16, 322-328. (11) Strausbauch, M. A.; Xu, S. J.; Ferguson, J. E.; Nunez, M. E.; Machacek, D.; Lawson, G. M.; Wettstein, P. J.; Landers, J. P. J. Chromatogr., A 1995, 717, 279-291. 10.1021/ac034648g CCC: $27.50 Published on Web 12/18/2003
© 2004 American Chemical Society
effects and consuming much less solvent than LLE were applied to LC16-21 and CE.18,22-29 However, the experimental arrangements were rather complicated and exhibited memory effects;30,31 moreover, the sensitivity improvement was insufficient. We have developed an efficient on-line preconcentration method using LPME prior to a CE run. A drop of the basic acceptor phase hanging at the capillary inlet tip was covered with a thin organic film. When the drop was placed in an acidic sample donor phase, acidic analytes were extracted to the acceptor phase through the organic film. The preconcentrated sample was then separated using a run buffer identical to the acceptor phase. A concentration ratio of 3 orders of magnitude was obtained for fluorescein with 30-min extraction. This simple technique consumes very little solvent, and no additional equipment is required. EXPERIMENTAL SECTION Chemicals. Sodium borate, fluorescein isothiocyanate (FITC), and 1-octanol were obtained from Sigma (St. Louis, MO). Hydrochloric acid was from Merck (Darmstadt, Germany). Sodium fluorescein was from Junsei (Tokyo, Japan). Deionized (12) Louter, A. J. H.; Vreuls, J. J.; Brinkman, U. A. T. J. Chromatogr., A 1999, 842, 391-426. (13) Vreuls, J. J.; Louter, A. J. H.; Brinkman, U. A. T. J. Chromatogr., A 1999, 856, 279-314. (14) Jones, H. K.; Stafford, L. E.; Swaisland, H. C.; Payne, R. J. Pharm. Biomed. Anal. 2002, 29, 221-228. (15) Pedersen-Bjergaard, S.; Rasmussen, K. E.; Halvorsen, T. G. J. Chromatogr., A 2000, 902, 91-105. (16) Ma, M. H.; Cantwell, F. F. Anal. Chem. 1998, 70, 3912-3919. (17) Ma, M. H.; Cantwell, F. F. Anal. Chem. 1999, 71, 388-393. (18) Rasmussen, K. E.; Pedersen-Bjergaard, S.; Krogh, M.; Ugland, H. G.; Grønhaug, T. J. Chromatogr., A 2000, 873, 3-11. (19) Zhu, L. Y.; Zhu, L.; Lee, H. K. J. Chromatogr., A 2001, 924, 407-414. (20) Zhu, L. Y.; Tay, C. B.; Lee, H. K. J. Chromatogr., A 2002, 963, 231-237. (21) Zhao, L.; Zhu, L.; Lee, H. K. J. Chromatogr., A 2002, 963, 239-248. (22) Pedersen-Bjergaard, S.; Rasmussen, K. E. Anal. Chem. 1999, 71, 26502656. (23) Pedersen-Bjergaard, S.; Rasmussen, K. E. Electrophoresis 2000, 21, 579585. (24) Halvorsen, T. G.; Pedersen-Bjergaard, S.; Rasmussen, K. E. J. Chromatogr., A 2001, 909, 87-93. (25) Halvorsen, T. G.; Pedersen-Bjergaard, S.; Reubsaet, J. L. E.; Rasmussen, K. E. J. Sep. Sci. 2001, 24, 615-622. (26) Zhu, L. Y.; Tu, C. H.; Lee, H. K. Anal. Chem. 2001, 73, 5655-5660. (27) Pedersen-Bjergaard, S.; Ho, T. S.; Rasmussen, K. E. J. Sep. Sci. 2002, 25, 141-146. (28) Ho, T. S.; Pedersen-Bjergaard, S.; Rasmussen, K. E. J. Chromatogr., A 2002, 963, 3-17. (29) Andersen, S.; Halvorsen, T. G.; Pedersen-Bjergaard, S.; Rasmussen, K. E. J. Chromatogr., A 2002, 963, 303-312.
Analytical Chemistry, Vol. 76, No. 3, February 1, 2004 855
Chino, CA) operated in the light-regulated mode at a power of 100 mW, and a wavelength of 488 nm was used as a light source for the laser-induced fluorescence detection. The laser beam was passed through a neutral density filter and focused with a Gradium GPX-5-12 lens (Ligthpath, Albuquerque, NM) into the capillary. Fluorescence emission from the sample was collected with a ×10 microscope objective (Edmund, Barrington, NJ), filtered through a 543-nm bass-pass filter (Melles Griot, Irvine, CA), and detected with an integrated photomultiplier tube (PMT; HC 120-01, Hamamatsu, Bridgewater, NJ). The power supply and PMT were interfaced to an IBM-compatible PC using an MIO-16-XE DAQ board (National Instruments, Austin, TX). Data acquisition and analysis were performed using custom-produced software developed with LabVIEW 5.0.1 (National Instruments).
Figure 1. Procedures of LPME/CE. (1) Octanol injection, (2) droplet formation at the capillary tip, (3) extraction, and (4) sample injection.
water was prepared with a NANOpure II purification system (Barnstead, Dubuque, IA). A run buffer was prepared by adjusting the pH of 20 mM sodium borate with 0.1 M NaOH to pH 9.50. A 1-mL aqueous sample solution of fluorescein and FITC (donor phase) was prepared by adding 10 µL of the standard solution in the run buffer to 990 µL of 0.10 M HCl (pH ∼1). Microextraction. Schematics of the extraction procedure are shown in Figure 1. The polyimide coating of a 50-cm fused-silica capillary of 75-µm i.d. (Polymicro Technologies, Phoenix, AZ) was removed to form a detection window 44 cm from the inlet end using concentrated hot sulfuric acid. The capillary was filled with the run buffer. The inlet end of the capillary was then placed in 1-octanol and the outlet end in a run buffer reservoir. Octanol was injected into the capillary by raising the inlet end by 12 cm, and the amount of injection was controlled by the injection time. After the inlet end was wiped with a piece of lint-free tissue to remove octanol on the outside of the capillary, the capillary was transferred to a capillary holder on a microcentrifuge tube containing an aqueous sample solution of 1.0 mL (donor phase). The microcentrifuge tube was lowered by 13 cm for 30 s to form a drop of the run buffer solution (acceptor phase) covered with octanol at the capillary tip. The tube was then placed in a jacketed water bath. The donor phase was stirred at a constant speed with a magnetic stirrer. During the extraction, the drop on the capillary tip was monitored with a video camera. After a desired extraction time, ∼12 nL of the drop was injected into the capillary by the force of gravity. The capillary was then removed from the capillary holder and placed into a run buffer reservoir for a normal CE run. CE Analyses. The capillary was conditioned each day by rinsing with 0.1 M NaOH for 5 min and with water for 5 min. Prior to each injection, the capillary was rinsed with the run buffer for 1 min. A CZE 1000R high-voltage power supply (Spellman, Hauppauge, NY) was used to apply a voltage of 10 kV across the capillary at room temperature. An argon ion laser (Omnichrome, (30) Pa´lmarsdo´ttir, S.; Thordarson, E.; Edholm, L.-E.; Jo ¨nsson, J. Å.; Mathiasson, L. Anal. Chem. 1997, 69, 1732-1737. (31) Thordarson, E.; Pa´lmarsdo´ttir, S.; Mathiasson, L.; Jo¨nsson, J. Å. Anal. Chem. 1996, 68, 2559-2563.
856
Analytical Chemistry, Vol. 76, No. 3, February 1, 2004
RESULTS AND DISCUSSION Basic Principles. LPME works through a series of two reversible extractions. The analyte in the aqueous donor phase (a1) is first extracted into the organic phase (o) and then backextracted into the aqueous acceptor phase (a2). At equilibrium, the enrichment factor (EF), which is defined as the ratio of the initial concentration of analyte (Ca1,i) in the donor phase to the equilibrium concentration (Ca2,eq) in the acceptor phase, can be represented as follows:16
EF ≡
Ca2,eq 1 ) Ca1,i D2/D1 + D2(Vo/Va1) + Va2/Va1
(1)
where the V represents the volume of the phase denoted by the subscript, and the distribution coefficients D1 and D2 are defined by
D1 ) Co,eq/Ca1,eq
(2)
D2 ) Co,eq/Ca2,eq
(3)
and
Let us consider a monoprotic analyte HA with pKa. Assuming that the neutral form is soluble only in aqueous phases, the distribution coefficient D can be given by
D)
K[H + ]aq
(4)
[H + ]aq + Ka
where K is the partition coefficient of the neutral form,
K ≡ [HA]o/[HA]aq
(5)
The ratio of distribution coefficients in eq 1 becomes
D2 [H+]a2 [H+]a1 + Ka ) D1 [H+] [H+] + K a1
a2
(6)
a
For a triprotic analyte H3A+ such as fluorescein and FITC, the
Figure 2. Electropherograms of (a) 3 nM FITC (1) and fluorescein (2) in a 20 mM sodium borate buffer and (b) 30 pM FITC and fluorescein enriched by 15-min LPME from a 1-mL sample solution at 35 °C using a 2-nL octanol layer. CE run buffer: 20 mM sodium borate (pH 9.5). Capillary: 50 cm (effective length 44 cm) × 75 µm i.d. Gravity injection for 5 s at 12 cm. Separation voltage: 10 kV. Fluorescence detection at 543 nm.
Figure 3. Preconcentration of fluorescein (O) and FITC (b) versus octanol volume. A 5-min extraction at 25 °C. Bars represent standard deviation (n ) 4).
expression for the ratio of distribution coefficients becomes
D2 ) D1
( )
[H+]a2 2 [H+]a13 + Ka1[H+]a12 + Ka1Ka2[H+]a1 + Ka1Ka2Ka3
[H+]a1 [H+]a23 + Ka1[H + ]a22 + Ka1Ka2[H+]a2 + Ka1Ka2Ka3 (7)
For pHa1 ) 1.0 and pHa2 ) 9.5, the distribution coefficient ratio for fluorescein becomes D2/D1 ) 9.1 × 10-8 using pKa1 ) 2.13, pKa2 ) 4.44, and pKa3 ) 6.36 for fluorescein.32 Then, with typical values of Va1 ) 1 mL, Vo ≈ 5 nL, and Va2 ≈ 100 nL, the enrichment factor can be approximated as
EF ≈ Va1/Va2 ≈ 104
(8)
Figure 2 shows the concentration effect of LPME. The electropherograms in Figure 2a and b were obtained by direct injection of analytes in a buffer of pH 9.5 and by injection followed by LPME for 15 min, respectively. An enrichment factor of 490 was obtained for fluorescein when the corrected peak areas () peak area/ migration time) were compared. Note that, for basic analytes, preconcentration using LPME should be easily achieved by reversing the pH of the donor and acceptor phases. Optimization of LPME. It was crucial to sustain the stability of the drop at the capillary tip. A few millimeters of polyimide coating were removed from the tip in order to prevent the drop from creeping up along the capillary outer wall. Since the drop was hanging on the capillary without any physical support, it became unstable under vigorous stirring conditions. To maintain a stable drop, control of the convection of the bulk solution was required. By stirring at 500 rpm with a 1-cm rod-shaped stirring bar, a drop of several nanoliters volume was well sustained in a vial containing an aqueous donor phase of 1.0 mL. The shape and (32) Diehl, H.; Markuszewski, R. Talanta 1985, 32, 159-165.
Figure 4. Preconcentration of fluorescein (O) and FITC (b) versus extraction temperature. A 5-min extraction using 2 nL of octanol. Bars represent standard deviation (n ) 4).
condition of the capillary tip were also important for the stability of the drop. When the capillary tip had scratches or was of irregular shape, the drop was easily destroyed during extraction. The enrichment factor depended on the volume and thickness of the organic phase. The octanol volume was estimated by using the Poiseulle equation.33 From the point of view of kinetics, the smaller the thickness, the shorter is the penetrating path. Then the time to reach equilibrium between the phases will be reduced. In our experiment, the volume of the octanol layer was changed from 2 to 20 nL by adjusting the octanol injection time. As illustrated in Figure 3, the amount of fluorescein extracted by LPME of 5-min duration increased with decreasing octanol volume. It was difficult to further reduce the volume of octanol, since the drop would become unstable. We investigated the temperature effect on the enrichment factor with a 1-mL sample solution, 2 nL of octanol, and 5-min extraction. With an increase in temperature from 20 to 35 °C, the enrichment factor increased by LPME from 62 to 240 for fluorescein and from 42 to 390 for FITC (Figure 4). This enrichment could be related to changes in the distribution equilibrium or to transfer kinetics. The measured values of the distribution coefficients of fluorescein, D1 and D2, were 320 ( 10 (33) Baker, D. B. Capillary Electrophoresis; John Wiley & Sons. Inc.: New York, 1995; pp 96-100.
Analytical Chemistry, Vol. 76, No. 3, February 1, 2004
857
and (2.0 ( 0.8) × 10-4 at 25 °C, and 280 ( 30 and (2.3 ( 1.0) × 10-4 at 35 °C, respectively. Thus, the term D2/D1 in eq 1 can be neglected at the both temperatures, suggesting that the increase in enrichment factor with temperature was mostly due to kinetic effects. At a temperature above 35 °C, a significant number of air bubbles were formed in the sample solution and these floated to the surface disturbing the drop at the capillary tip. The enrichment factor was also affected by the extraction time. When the extraction time was increased from 5 to 30 min with 7 nL of octanol, the enrichment factor for fluorescein increased as 42 ( 5 (5 min), 210 ( 60 (15 min), and 1200 ( 500 (30 min). With a longer extraction time, the reproducibility became poorer and the drop tended to become detached from the capillary tip. The relative standard deviations for corrected peak area of fluorescein with 15-min extraction were ∼30%, probably due to the uncertainties in the manual control of the droplet formation. We attributed the discrepancy between the actual enrichment factors and the theoretically expected values from eq 8 to kinetic effects. However, the high efficiency attained with this short extraction time, compared to other reported values,22 was due to the thin wall of the organic phase. Advantages of LPME. There are many advantages of using our LPME scheme for CE. No additional instrumentation is needed. Since the forward- and back-extractions are carried out simultaneously in one process, multiple manipulations such as solvent addition, phase separation, and solvent evaporation are not required. Cross-contamination can be easily avoided by using
858
Analytical Chemistry, Vol. 76, No. 3, February 1, 2004
fresh organic solvent for each extraction. The small drop size reduces the extraction time and also minimizes the consumption of organic solvent. In addition, a sample cleanup effect was observed in LPME. High salt solutions usually cause destacking and poor separation efficiency in CE. Even when the donor phase contained 0.1 M NaCl in addition to 0.1 M HCl, the enrichment of fluorescein was accomplished without a decrease in peak efficiency or enrichment factor, implying that this method could be applied to physiological samples in highly saline matrixes. CONCLUSIONS We applied LPME as a convenient preconcentration method for CE. By adjusting the pH of the aqueous solutions and using an octanol layer of 7 nL, enrichment of fluorescein by a factor of 3 orders of magnitude was accomplished with 30-min extraction. Compared to LC or GC, CE was most suitable for applying liquidphase extraction since the small volume drop matched the small injection volume in CE. An LPME study using a commercial CE instrument is in progress. ACKNOWLEDGMENT This work was supported by the R&D Program for Fusion Strategy of Advanced Technologies, IMT 2000, and BK21 program of Korea. Received for review June 14, 2003. Accepted November 13, 2003. AC034648G
