Anal. Chem. 2003, 75, 3656-3659
Correspondence
CE in a Nonuniform Capillary Modulated by a Cylindrical Insert, and Zone-Narrowing Effects during Sample Injection Alexander V. Stoyanov, Zhen Liu, and Janusz Pawliszyn*
Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1
The electrophoretic behavior of an analyte in a capillary consisting of two parts of different cross section has been investigated. Modulation of the cross-sectional area of the separation channel has been achieved by inserting a cylindrical fiber different distances into the capillary. It was shown that the zone injected into the capillary part with smaller cross section could be moved using electromigration into the wider part of the capillary with zone compression. As we observed, the zone narrowed longitudinally in the wide part of the capillary in accordance with the ratio of the electric field strength in the two parts of the capillary. The concentration of plug introduced into the capillary by electroinjection can be increased by use of low-conductivity sample buffer. Efficient introduction of extracted analytes desorbed from an SPME fiber to the separation channel was achieved using this approach. Thermoinduced effects caused by temperature increase in the narrow part of the capillary and their influence on sample stacking are analyzed. Possible applications of the effect observed to the sample introduction optimization are also discussed in this study. Use of a nonuniform channel in electrophoresis is a means of varying the electric field strength along the separation path.1,2 The appropriate electric current density change can influence other conditions (temperature, pH), which also play an important role in the process. The other means to vary the electric field strength is to vary the separating media properties,3-6 which can be applied in a straight channel. Smooth variation of the cross section of the electrophoretic channel is required to obtain a smooth electric field function. By slight progressive modification of the cross section of the electrophoretic capillary, one obtains an electric field gradient. This effect can be used in IEF, which is conducted with the help of thermogradients, provided the current density drop and the * Corresponding author. E-mail:
[email protected]. (1) Pawliszyn, J.; Wu, J. J. Microcolumn Sep. 1993, 5, 397-401. (2) Slais, K.; Freidl, Z. J. Chromatogr. 1994, 684, 149-161. (3) Luner, S.; Kolin, A. Proc. Natl. Acad. Sci. U.S.A. 1970, 66, 893-903. (4) Lundhal, P.; Hjerten, S. Ann. N. Y. Acad. Sci. 1973, 209, 94-11. (5) Greenlee, R. D.; Ivory C. F. J. Biotechnol. Prog. 1998, 14, 300-309. (6) Ross, D.; Locascio, L. E. Anal. Chem. 2002, 74, 2556-2564.
3656 Analytical Chemistry, Vol. 75, No. 14, July 15, 2003
chamber design are suitable for generating a sufficient temperature difference.1,7-10 In contrast, a step change in the crosssectional area results in a sharp change in the field strength, and for a stepwise increase in cross-sectional area, the corresponding decrease in electric field strength results in sample stacking. A separation channel composed of a few different parts, each of constant form, can be used for sample introduction, detection, and multistep analytical separations in a microarray, etc. The results described in this paper illustrate this effect achieved by insertion of a long object (e.g., an SPME microfiber) into the separation capillary. EXPERIMENTAL SECTION Apparatus. Whole-column imaging detection by UV absorbance was conducted with an iCE280 CIEF instrument (Convergent Bioscience Ltd, Toronto, Canada) at a fixed wavelength of 280 nm. A short fused-silica capillary (5.5 cm long) with an i.d. of 100 µm, internally coated with fluorocarbon (J&W Scientific, Folsom, CA), was assembled in a cartridge format (Convergent Bioscience Ltd.). The entire process of capillary conditioning, sample injection, data collection, and processing was controlled by means of a personal computer, and the electropherogram absorbance signal was recorded versus the distance to the anode. Materials and Chemicals. Optical fibers with 50- and 61.5µm cores (FHP050055065 and FVP60072082) were purchased from Polymicro Technologies Inc. (Phoenix, AZ). Synthetic pI markers and buffer chemicals were obtained from Bio-Rad (Hercules, CA). Water purified by use of an ultrapure water system (Barnstead/ Thermolyne, Dubuque, IA) was used for all solutions. Procedures. The fiber was inserted different distances into the capillary, and the capillary was filled with running buffer (100 mM phosphate, pH 2.5, BioRad). The sample (BioMark synthetic pI marker, pI 10.0) was then injected electrokinetically with the injection time being selected to achieve complete displacement of the liquid in the first part of the capillary (i.e., that containing the inserted microfiber). The electrode reservoirs were then (7) Fang, X.; Tragas C.; Wu J.; Mao, Q.; Pawliszyn, J. Electrophoresis 1998, 19, 2290-2295. (8) Fang, X.; Adams, M.; Pawliszyn, J. Analyst 1999, 124, 335-341. (9) Pawliszyn, J.; Ciu, Y.-P.; Semenov, S. N. Instrum. Sci. Technol. 2000, 28, 281-301. (10) Huang, T.; Pawliszyn, J. Electrophoresis 2002, 23, 3504-3510. 10.1021/ac026395h CCC: $25.00
© 2003 American Chemical Society Published on Web 06/07/2003
Table 1. Sample Plug Narrowing for Two Cylindrical Inserts of Different Diameter compression factor
Figure 1. Microcartridge with microfiber inserted at the anode end. By inserting a microfiber of an outer diameter close to the inner diameter of the capillary it is possible to achieve a large increase in the electric field strength in the first part of the capillary.
washed and filled with the desired separation buffer, and the electrophoretic run was performed. RESULTS AND DISCUSSION The initial zone width is important in CZE. If sample concentration is insufficient to enable sensitive detection, several on-line enrichment procedures can be used to remedy the problem.11-14 The simplest electrophoresis-based techniques entail creation of a special conductivity profile, which allows for a higher electric field strength in the sample zone. Different concentration mechanisms (e.g., CE- or ITF-based) can be used in these procedures. A similar effect of electric field enhancement can be achieved by use of a stepwise change in the cross-sectional area. In our experiments we modulated the cross section of the separation channel by inserting a cylindrical microfiber (Figure 1). Sample was injected electrokinetically at 500 V, and the duration of the voltage pulse was adjusted to achieve complete filling of the first part of the capillary (up to the end of the microfiber). The initial starting zone was rather wide and corresponded to the length to the inserted fibersapproximately half of the capillary, taking into account the “dead” volume. It was then effectively compressed to the zones showed as trace II in Figure 2A and trace IV in Figure 2B. This transformation occurred in the same proportion as is expected from the cross-sectional area ratio. In the case of two coaxial cylinders, the ratio R ) S2/S1 (S1 and S2 are the areas of the cross sections for narrow and wide parts, respectively) is related with the diameters
R ) D2/(D2 - d2)
(1)
where D is inner the diameter of the capillary and d is the outer diameter of the microfiber inserted. Assuming the conductivity is constant, the electric field increase in the narrow part (E1/E2) is defined by S2/S1, and the initial zone length should be narrowed in approximately the same proportion. Figure 2 illustrates the zone width decreasing by a factor of 3 for the 82-µm-o.d. insert and by a factor of 2, approximately, for the 65-µm-o.d. insert; for these examples R ) 3.1 and 1.7, respectively; see Table 1. The concentration profiles were obtained with 40-s time intervals. (11) Johansson, J.; Witte, D.; Larsson, M.; Nilssom, S. Anal. Chem. 1996, 68, 2766-2770. (12) Stroink, S.; Paarlberg, E.; Waterwal, J.; Bult, A.; Underberg, W. Electrophoresis 2001, 22, 2374-2383. (13) Lichtenberg, J.; Verpoorte, E.; de Roij, N. F. Electrophoresis 2001, 22, 258272. (14) Burgi, D.; Chen, R. Anal. Chem. 1992, 64, 489A-493A.
insert diameter (d), µm
experimental
theory according to eq 1
65 82
1.8 3.1
1.7 3.1
The effect observed can be combined with methods traditionally used for sample preconcentration. When low-conductivity buffer was used,15 it was possible to achieve an increase in the concentration of the sample plug introduced (up to 10 times), although the change in the peak width was less evident. We must emphasize that the effect at low current densities described above does not itself provide any increase in the concentration of the sample introduced, because the volume of the sample zone should remain constant, and the narrowing of the sample plug is because of its change in shape. Indeed, for the component analyzed, the total flux (F) is the same in each electrophoretic channel of different cross-sectional area:
F ) SvC ) const
(2)
where C is the concentration and v is the velocity of electrophoretic migration. The velocity, v, is proportional to the electric field strength, E, which is, in turn, inversely proportional to the cross-sectional area; thus, the product Sv remains constant. This is the condition of sample-plug volume constancy when the separation channel cross section is changed and it follows from formula 2 that the sample concentration remains constant. Although the concentration does not change as the zone changes its form, this simple approach to zone narrowing opens the way to many important applications in which the separation is started from an “initially wide” zone when it is necessary. It is much easier to introduce a long plug to a narrow channel then a short plug to a wider channel. Unlike pressure-driven flow, in a nonuniform capillary or channel, much more uniform samplezone transformation can be achieved in electrophoretic zone migration. For example, when working with a coated SPME fiber one can insert a microfiber into the capillary, obtaining a rather wide starting zone, but on application of the electric field, the initial zone can effectively be narrowed at the end of the microfiber. The size of this effect depends on the relative size of the microfiber insertedsto achieve substantial zone narrowing, the difference (D2 - d2) should be appropriately small. The plug form change could also be very useful to facilitate coupling of two or more techniques using different channel dimensions (for example, OTLC and CE). Additionally, this approach can be used as an alternative to a standard “cross” injection in microfluidic systems. Also, the increase in the signal due to the increase in the path length for absorbance detection in the wide part of the separation channel could provide an opportunity to increase the sensitivity of detection. The effect observed can be combined with methods traditionally used for sample enrichment such as SPE, SPME, or electro(15) Palmer, J.; Burgi, D.; Munro, N.; Landers, J. Anal. Chem. 2001, 73, 725731.
Analytical Chemistry, Vol. 75, No. 14, July 15, 2003
3657
Figure 2. Behavior of a long sample plug in CZE when the initial part of the separation channel has a smaller cross section. The sample, Biomark synthetic pI marker (pI ) 10.0), was initially concentrated in the narrow part of the capillary and the total plug length was approximately half of the length of the capillary. The running buffer was 100-mm phosphate (BioRad); the sample was dissolved in the same buffer. The diameters of the cylindrical inserts were 82 (A) and 65 µm (B), and the runs were performed at 500 and 1500 V, respectively. The concentration profiles were obtained after 0, 40, 80, 120, and 160 s (curves I, II, III, IV, V). Scanning distance is 5 cm.
phoretic focusing approaches. For example, using the technique of samples introduced in low-conductivity buffer,14 it was possible to achieve an increase in the concentration of the sample plug injected (up to 10 times). For SPME coupled with capillary electrophoresis, the SPME microfiber can be inserted directly into the capillary.16 Under these conditions, the rate of analyte desorption is limited, to some extent, by the progressive increase in analyte concentration near the microfiber surface. Application of the electric field substantially promotes desorption, because the desorbed analyte is removed and thus the concentration gradient is increased. Although desorption of analyte from the microfiber surface continues during the electrophoretic run, resulting in zone broadening, the effect described above makes it possible to achieve sample concentration in the vicinity of microfiber tip. Figure 3 illustrates the analyte concentration distribution in the open part of the channel after electrophoresis for 60 s. Here, the experimental set was modified and the laser-induced fluorescence signal was collected. The presence of the fiber in the separation channel does not have a negative impact on the zone width. This phenomenon can be used effectively to produce a narrow injection plug when coupling other external sampling/sample preparation microdevices, e.g., a microneedle or a micro SPE cartridge, to the separation channel. In our consideration so far, the effect of anticipated temperature difference in the two parts of the capillary has been neglected. This assumption is justified by the low current density in our experiments. Nevertheless, it is useful to analyze the more general case of some temperature difference between the two parts of the capillary and its possible contribution to the stacking effect. As discussed in ref 1, higher heat production per unit volume in the narrow part of the capillary results in higher temperature compared to the wider part. We investigated zone transformation in different buffers with progressively increasing voltages (500, 1500, and 3000 V). Additionally, to achieve higher heat dissipation in the narrow part of the capillary, the total current density was (16) Whang, C. W.; Pawliszyn J. Anal. Commun. 1998, 35, 353-356. (17) Huang, T.; Pawliszyn J. Analyst 2000, 125, 1231-1233.
3658 Analytical Chemistry, Vol. 75, No. 14, July 15, 2003
Figure 3. Sample distribution in the capillary after sample introduction with SPMI fiber. The two-dimensional concentration distribution of analyte (dansyl-phenylalanine, Molecular Probes, Eugene, OR), initially adsorbed on an SPME fiber, after desorption and 60 s of electrophoretic run (A). The laser-induced fluorescent signal was collected. The upper panel shows the fluorescence intensity scale. (B) Line I shows the concentration along the capillary axis at the same time (60 s). Line II is the concentration profile before application of voltage. The optical fiber used both for analyte adsorption and for axial illumination was coated, as described elsewhere.16 The sample was adsorbed at the end of the microfiber tip (0.9 cm). The running buffer was 10 mM sodium phosphate, pH 8.8. The run was performed at 500 V.
modulated by varying the length of the insert. Only small differences in zone shape were observed with current density increase.This result agrees with our expectations, according to which even considerable temperature difference is not able to produce by itself a substantial concentration effect.
Indeed, as temperature rises, the viscosity decreases, thus resulting in mobility increase, according to
µ ) q/6πηa
(3)
where q is the charge of the analyte molecule, η the buffer viscosity, and a the solute radius. On the other hand, one could expect the same mobility changes for all charged species; the latter results in an appropriate conductivity (σ) change
σ∼
∑q µ
i i
(4)
An electric field will decrease in accordance with the Ohm’s law
E ) j/σ
(5)
here, j is used for current density because of increased conductivity. Since the increased mobility is counteracted by the decreased electric field, we may conclude the net effect of higher temperature on analyte migration velocity should be close to nil at the first approximation. Other effects, such as thermal expansion and diffusion coefficient increase, can contribute to zone defocusing. To take advantage of thermoinduced stacking effects, one needs to take into account that temperature impact on analyte mobility has a more complicated character since the change of mobility can be both positive and negative for different species present in the solution. Therefore, the focusing effect can be obtained by optimizing the composition of the buffer. To describe (18) Pawliszyn, J. Anal. Chem. 1988, 60, 2796-2801.
this effect correctly, one should also consider the change in ionization constants with temperature, for both analyte and buffer components. Although, taking into account the rather small value of dissociation constant temperature coefficients, one should not expect typically a significant temperature influence on the resulting analyte mobility, unless the buffer system is especially designed to emphasize this effect.1,3 Similarly, high sample concentration (comparable to the concentration of buffer) could provide some small focusing/defocusing effect that depends on acid-base properties of both the sample and the buffer. In this study, we investigated electrophoretic zone behavior in a nonuniform separation channel. We observed zone transformation with the length decreasing according to an increase in cross-sectional area ratio. Such transformation allows us to introduce sample more effectively in cases when the initial starting zone is too long. The results of this work studying analyte behavior in a nonconstant free cross-sectional area separation channel might also be important for optimizing the sample introduction in other separation and detection18 methods and for sample manipulation in the case of multidimensional separations. ACKNOWLEDGMENT The authors thank the Natural Sciences and Engineering Research Council (NSERC), Canada and Convergent Bioscience for financial support.
Received for review December 4, 2002. Accepted April 23, 2003. AC026395H
Analytical Chemistry, Vol. 75, No. 14, July 15, 2003
3659
