A Microchip Device for Enhancing Capillary Zone Electrophoresis

Oct 23, 2012 - Department of Chemistry, University of Wyoming, Laramie, Wyoming 82071, United States. Anal. Chem. , 2012, 84 (22), pp 10058–10063...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

A Microchip Device for Enhancing Capillary Zone Electrophoresis Using Pressure-Driven Backflow Ling Xia and Debashis Dutta* Department of Chemistry, University of Wyoming, Laramie, Wyoming 82071, United States ABSTRACT: A reduction in the electroosmotic flow (EOF) is often desirable in glass microchannels for realizing high resolutions in capillary zone electrophoresis (CZE). While static and dynamic coatings have been commonly employed to accomplish this goal, such chemicals can introduce unwanted interactions of the analyte molecules with the separation medium and/or channel surface. In this article, we report a microfluidic device that can enhance the resolving power of CZE analysis by generating a pressure-driven backflow in the separation channel. This backflow was generated in our current work by fabricating a shallow segment (0.5−4 μm deep) downstream of the separation duct (5 μm deep) and applying an electric field across it. A mismatch in EOF transport rate at the interface of this segment was shown to yield a pressure-gradient that counteracted electroosmosis and diminished the net fluid flow in the separation conduit by nearly an order of magnitude. Although the resulting pressure-driven backflow also somewhat increased the band broadening in the analysis channel, overall it allowed us to separate an amino acid mixture with an 8-fold higher resolution. The microchip device presented here is particularly suitable for miniaturization of the CZE method and may be easily integrated into other analytical procedures making it an attractive module for lab-on-a-chip applications.

C

methods. Static coatings, on the other hand, are particularly unsuitable for the miniaturization of the CZE method due to increasing difficulties in applying them to the channel walls and the enhanced interaction of the solute molecules with the channel surface as the size of the analysis column is reduced to the micro/submicrometer length scale. Hayes and co-workers reported for the first time, the application of external voltages at the channel walls for local control of the zeta potential, and therefore the EOF, in a dynamic fashion.18 Although this approach overcomes some of the disadvantages of the static/ dynamic coatings, it offers somewhat limited control over the range of realizable electroosmotic transport rates and tends to complicate the instrumentation aspects of the CZE method. The use of a pressure-driven flow to counteract electroosmosis in CZE analysis presents an alternate approach to reducing the separation length of this technique.19,20 In spite of being relatively simple to implement, this strategy unfortunately introduces additional band broadening due to the parabolic nature of the pressure-driven flow profile21 and thereby compromises the resolving power of the CZE method. Interestingly, such sample dispersion scales directly with the lateral dimensions of the analysis channel22−24 and can therefore be reduced by miniaturizing the CZE columns. The ability to generate pressure-driven backflows with high precision and dynamic control in shallow microchannels however presents a significant challenge25 that has prevented

apillary zone electrophoresis (CZE) is an important scientific tool capable of liquid-phase separation of charged samples in a rapid and reliable manner.1,2 Over the past two decades, the significance of this method has grown even further with the emergence of lab-on-a-chip systems which have allowed its easy integration to various sample preparation and analysis procedures.3−5 Moreover, the miniaturization of CZE on the microfluidic platform has permitted the realization of higher separation speeds and efficiencies using this technique, with smaller sample sizes than previously possible.6−8 Unfortunately, the availability of limited footprint area on planar microchip devices has restricted the applicability of microfluidic CZE to relatively simple separations. While several strategies have been reported in the literature for accommodating long analysis channels in compact microchip systems,9−13 there is a need for developing approaches that would allow difficult separations to be realized in relatively short microfluidic ducts. Since the advent of the CZE method, a variety of strategies have been identified to shorten the separation length of this technique. Among them, the use of surface and dynamic coatings to reduce electroosmotic flow (EOF) in the analysis channel has been commonly employed in the literature.14−16 While dynamic coatings are easier to apply and can withstand higher pH values than the static ones, they tend to be less stable than the latter.17 Moreover, these chemicals can introduce unwanted interactions of the analyte molecules with the separation medium affecting the resolution of the assay as well as interfere with the direct integration of the CZE technique to several detection and/or downstream analysis © 2012 American Chemical Society

Received: September 2, 2012 Accepted: October 23, 2012 Published: October 23, 2012 10058

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063

Analytical Chemistry

Article

Figure 1. (a) Schematic of the microchip device used in the present work. Notice that the magnitude of the pressure-driven flow in segments A, B, and C is expected to be about a third of that in segment D in this design. (b) Gated injection scheme employed for introducing sample plugs into the separation channel of the microchip device shown in (a). The fluorescence images presented above were obtained using a 10 μM rhodamine B sample prepared in a 1 mM sodium tetraborate buffer applying φ0 = φ1 = 1 kV in stage III of the injection process.

created through Fineline Imaging Inc. (Colorado Springs, CO). The channel width in our device was chosen to be 100 μm for the entire microfluidic network, and the separation column in it (segments D) was made 3 cm long. After completion of the photopatterning process, the photoresist layer was cured in microposit developer MF-319 (Rohm and Haas) and the chromium layer removed along the channel network with a chromium etchant (Transene Inc.). The shallow segment needed for pressure generation was created by first etching the entire microchip to the depth of the shallow region (0.5−4 μm) with buffered oxide etchant (Transene Inc.), and then manually covering up a 1 mm long region 2.5 cm downstream of the injection cross (the location of the shallow region) with a layer of photoresist. After drying this photoresist layer in a convection oven at 80 °C for 20 min, the remaining fluidic network was further etched to a final depth of 5 μm. Following this step, four access holes were punched at the channel terminals using a microabrasive powder blasting system (Vaniman Inc.). Finally, the microfluidic network was sealed off by bringing a cover plate in contact with the bottom substrate in deionized water and allowing the two plates to bond at 550 °C for 9 h.33 Device Operation. The reported microchip was prepared for an experiment by rinsing the entire fluidic network with 0.1 N sodium hydroxide followed by deionized water and the appropriate buffer solution for 20 min each. Analyte injection into the separation channel (segment D) was accomplished by placing the sample solution in reservoir 1 while filling the remaining reservoirs with the corresponding buffer. The application of high voltages to reservoirs 1 and 2 (1 kV each) under these conditions then allowed the realization of a sample flow profile that was suitable for analyte injection (see Figure 1b, stage I). In order to introduce sample slugs into segment D of our device, a gated injection scheme was used.34 This scheme was implemented by electrically grounding reservoir 2 for 0.3 s, during which the sample was allowed to fill up the entire injection cross (see Figure 1b, stage II). Following this step, the electric potentials at reservoirs 1 and 2 were switched back to

the use of this transport mode for improving the performance of microfluidic CZE. Although several devices have been described in the literature for performing liquid phase chromatography in microchannels using an on-chip pressure generation capability,26−31 there is no report of utilizing these micropumps for enhancing CZE assays. In this article, we present a simple approach to introducing a steady pressure-driven backflow in microfluidic CZE columns and thereby improve the resolving power of this analysis technique. This backflow was generated in our current work by fabricating a shallow segment downstream of the separation channel and applying an electric field across it. A mismatch in EOF transport rate at the interface of this segment was shown to yield a pressure gradient30 that counteracted electroosmosis and diminished the net fluid flow in the separation conduit by nearly an order of magnitude. While the resulting pressuredriven backflow also somewhat increased the band broadening in the analysis channel, the reported device was shown to improve the resolution of CZE separations by a factor of 6−8. The microchip design presented here for enhancing electrophoretic analysis using pressure-driven backflow offers three major benefits: (1) it omits the use of an external pump for pressure generation simplifying its fabrication and operation; (2) it can readily be integrated into any detection and/or downstream analysis method; and (3) it is suitable for further miniaturization of the CZE technique to the submicrometer length scale.

■

EXPERIMENTAL SECTION Device Fabrication. The microchip device shown in Figure 1a was fabricated using a bottom substrate and cover plate made from borosilicate glass (Telic Company, Valencia, CA). While the purchased cover plates had both of their faces unprotected, the bottom substrates came with a thin layer of chromium and photoresist laid down on one of their surfaces. The fabrication process for the microchips was initiated by photolithographically patterning32 the desired channel layout on the bottom substrate using a custom-designed photomask 10059

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063

Analytical Chemistry

Article

Figure 2. (a) Variation in fluid transport rate as a function of the operation voltage (φ0) in our CZE device. (b) Measured fluid velocity as a function of the depth of the shallow region in segment D at φ0 = 2 kV. (c) Estimated electric field in segment D of our CZE device at φ0 = 2 kV. (d) Estimated pressure-driven backflow velocity in the separation channel at φ0 = 2 kV. The error bars included in (b) and (d) are based on 5 measurements.

magnitude of Unet is reduced by nearly a factor of 4 upon introduction of a 1 μm deep shallow segment in the separation channel. The observed reduction in flow velocity primarily occurs due to generation of a pressure gradient at the interface of the shallow region as has been demonstrated previously.30 The resulting pressure-driven flow counteracts electroosmosis in segment D and increases proportionally with the electric field in it,30 yielding the linear variation between Unet and φ0. On the other hand, the rise in fluid velocity with a decrease in ionic strength of the background electrolyte occurs due to an increase in the zeta potential at the channel walls.35,36 In Figure 2b, we have presented the observed variation in Unet as a function of the depth in the shallow region (ds) in segment D for φ0 = 2 kV. The figure shows a sharp decrease in this measured velocity for all operating voltages and background electrolytes when ds is reduced below 3 μm. For ds = 0.5 μm, this quantity is seen to decrease by about a factor of 8 compared to its corresponding value in a separation channel without the shallow segment. It must be noted that the observed decrease in Unet occurs both due to a reduction in the actual electric field as well as an increase in the pressure-driven backflow in segment D. The contributions from each of these factors have been estimated in Figure 2, panels c and d, respectively, which show the latter effect to be the dominant one in our experiments. For example, in the device with a 0.5 μm deep shallow region, it is estimated that the electric field in the separation channel is reduced by at most 20%, whereas the pressure-driven backflow in it decreases the net fluid transport rate by a factor of 7−9. The electric field values presented in Figure 2c were evaluated based on current−voltage measurements made across the different segments of our microfluidic network. The electrical resistances thus estimated for the 4 channel segments were later used in calculating the voltage drop across the separation channel through application of

the operation voltage to be used in the experiment, which also forced a fraction of the sample that had filled up the injection cross to be swept into segment D (see Figure 1b, stage III). The sample plugs thus injected into the separation channel were detected 1.5 cm downstream from the injection cross via the laser-induced fluorescence technique. For these measurements, a laser beam (125 mW argon ion tunable laser, Melles Griot) of wavelength 488 nm (in the experiments involving FITC-labeled amino acids) or 514 nm (in the experiments involving resorufin and rhodamine B) was used to excite the dye molecules and the resulting fluorescence signal collected using a photomultiplier tube (Hamamatsu Photonics) after passing it through an appropriate (514 nm long pass for the experiments involving FITC labeled amino acids and 585 ± 20 nm bandpass for the experiments involving resorufin/rhodamine B) optical filter (Semrock Inc.). The sequence of voltages applied to the different microchannel terminals during the sample injection/separation stages and the process of data acquisition with the photomultiplier tube were controlled using a LABVIEW program. The peak shapes were analyzed in our work using numerical data analysis tools available in MATLAB.

■

RESULTS AND DISCUSSION Pressure-Driven Backflow Measurements. In order to quantitatively understand the performance of the reported device, we initiated our study by investigating the effect of the shallow region on the fluid transport rate (Unet) in its separation segment. These measurements were made by introducing 10 μM samples of a neutral tracer, rhodamine B, into segment D and determining their transit times between the injection and the detection points. Figure 2a shows that the quantity, Unet, increases linearly with the operating voltage (φ0) and at a rate that depends inversely on the ionic strength of the background electrolyte. Moreover, for a given value of φ0, the 10060

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063

Analytical Chemistry

Article

Kirchhoff’s and Ohm’s laws. Assuming the fluid conductivity in the deep portion of segment D to be identical to that in segments A, B, and C, we then evaluated the Ohmic resistance of this region based on its linear relationship with the ratio of the length to the cross-sectional area of the channel segment. This quantity was then subtracted from the electrical resistance of segment D to estimate the corresponding quantity for the shallow region in our device. Finally, the voltage drop and, subsequently, the electric field in the deep portion of segment D were evaluated by applying the concept of a voltage divider to the current flow through the separation channel. The Ohmic resistance measurements described above indicate a higher electrical conductivity of the fluid in the shallow region particularly for low ionic-strength electrolytes as has been observed in our previous work, yielding greater pressure gradients and larger electric fields in the separation channel.30,37 The pressure-driven backflow reported in Figure 2d was estimated by subtracting the electroosmotic velocity in segment D from the Unet values presented in Figure 2b. Electroosmotic mobility measurements in glass channels without a shallow region and the electric field values included in Figure 2c were used to determine the electroosmotic velocity in the above calculations. A drawback of the strategy reported here to reduce the fluid flow rate in CZE systems is its tendency to increase the broadening of sample bands. Such broadening occurs due to the nonuniformity in the pressure-driven flow profile and is known to scale directly with the shortest lateral dimension of the analysis column.21−23 However, because diffusion across microchannel cross sections tends to be rapid, flow dispersion in the reported microfluidic CZE device is expected to be small. This point has been established in Figure 3 by measuring the

the stronger dependence of R on the average analyte velocity compared to that on H as suggested by the expression, R = 1 Δu√L/(4u√H). In the above expression, Δu and u̅ refer to ̅ the differential and average velocities for the two analyte bands being separated, while L denotes the separation length in the system. Dye Separation. Having established the flow characteristics in the reported CZE device, we proceeded to determine the effect of the various operating parameters on its resolving power for a mixture of two fluorescent dyes, rhodamine B (electro-neutral) and resorufin (anionic). In Figure 4a, we have presented a series of electropherograms obtained from these experiments using 10 μM dye samples prepared in a 1 mM sodium tetraborate buffer for different depths of the shallow region in segment D. The operating voltage (φ0) used in this study was 2 kV, and the analytes were detected 1.5 cm downstream from the injection point. These plots clearly demonstrate an increase in the resolving power of our CZE device as the depth of the shallow region in its separation segment is reduced. In Figure 4b, we have quantitated these results showing an improvement in the separation resolution (R) for the dye mixture by about a factor of 6 using a 0.5 μm deep shallow segment under the chosen experimental conditions. The quantity R in this figure was estimated as R = Δt/(4σ̅), where Δt and σ̅ correspond to the differential migration time and average standard deviation (in time units) for the rhodamine B and resorufin peaks in the electropherogram. The figure also shows a very different effect of the operation voltage (φ0) on R, as the depth of the shallow region (d) is reduced in segment D. While the resolution between the peaks is seen to monotonically increase with an increase in φ0 for devices with ds ≥ 2 μm, this quantity reaches a maximum around φ0 = 2 kV when ds = 0.5 μm. Because both the pressuredriven and electrokinetic transport rates for the analyte species scale linearly with the operation voltage in our system, the relationship R = Δu√L/(4u√H) predicts the maximum in the ̅ separation resolution curve to be dictated by the variation in H with φ0. In particular, R is expected to reach its peak value around the same operation voltage at which the plate height reaches a minimum. Figure 4b, however, shows that the resolving power of our device is maximized at an operation voltage which is somewhat greater than expected based on the relationship presented above. As the reason for this discrepancy is currently unclear to us, we have chosen not to present any design rules for predicting the location of the maximum in the separation resolution curve in our system. Importantly, Figure 4b establishes that the optimum separation condition (i.e., φ0 = 2 kV) for our best performing device (i.e., the one with ds = 0.5 μm) may be readily realized using our current experimental setup. Amino Acid Separation. Based on our findings from the dye separation experiments, we applied our CZE device to enhancing the electrophoretic analysis of a 10 μM FITClabeled amino acid mixture (see Figure 5). The analytes chosen for this study included arginine, glutamic acid, and glutamine, all of which were observed to be anionic after conjugation to FITC in a 1 mM sodium tetraborate buffer (pH 9.1). As was the case for the dye sample, the separation resolution for the amino acid mixture was observed to improve significantly in our CZE device upon introduction of a 0.5 μm deep shallow region in segment D. In fact, for the chosen amino acid sample, a nearly 8-fold improvement in the separation resolution was realized for an optimum operating voltage of 2 kV. Keeping in

Figure 3. Plate height for the rhodamine B peaks in a 1 mM sodium tetraborate buffer measured 1.5 cm downstream of the injection cross. The error bars included in the figure are based on 5 measurements.

plate height (H) of rhodamine B peaks in our separation channel. The data presented in the figure shows that for a device without a shallow region, H decreases with an increase in φ0 as expected based on the diffusive broadening of the analyte bands. Upon introduction of the shallow segment, this dependence reverses particularly at large operating voltages due to flow dispersion arising from the pressure gradients in the system. For ds = 0.5 μm and φ0 = 3 kV, the plate height is seen to increase by as much as a factor of 3 over its corresponding value in channels without a shallow region. In spite of this increase in band broadening, we expect a significant improvement in the separation resolution (R) of our CZE device due to 10061

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063

Analytical Chemistry

Article

Figure 4. (a) Effect of the depth of the shallow region in segment D of our CZE device on its resolving power. The electropherograms presented here were obtained using a 1 mM sodium tetraborate solution as the background electrolyte and an operation voltage of 2 kV. (b) Measured separation resolution for the rhodamine B and resorufin peaks shown in (a) as a function of the various operating conditions. The error bars included in (b) are based on 5 measurements.

Figure 5. (a) Enhancement in the CZE analysis of an FITC-labeled amino acid sample using the reported microfluidic device. The symbols R, Q, and E in the electropherograms presented above refer to the amino acids arginine, glutamine, and glutamic acid, respectively. (b) Measured separation resolution for the amino acid peaks shown in (a). The symbols ● and ⧫ here refer to the separation resolutions between the FITC-R and FITC-Q and the FITC-Q and FITC-E peaks, respectively. The open and filled symbols here correspond to measurements made in separation channels without a shallow region and with a 0.5 μm deep one, respectively. The error bars included in (b) are based on 5 measurements.

amino acid sample as compared to those realized for the dye mixture were not surprising. For both samples, however, the enhancement in the resolving power of the CZE analysis was accomplished at the cost of a lower separation speed. Interestingly, the separation time for our CZE device may be

mind that all of the FITC-labeled amino acids were anionic in nature and underwent electrophoresis in a direction opposite to the EOF, their average velocity (u̅) was expected to be slower compared to the value of u̅ in the case of the dye mixture. In this situation, the higher separation resolutions obtained for the 10062

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063

Analytical Chemistry

Article

(2) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z. H.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895−897. (3) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem. 2002, 74, 2637−2652. (4) Vilkner, T.; Janasek, D.; Manz, A. Anal. Chem. 2004, 76, 3373− 3385. (5) Dittrich, P. S.; Tachikawa, K.; Manz, A. Anal. Chem. 2006, 78, 3887−3907. (6) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 1114−1118. (7) Jacobson, S. C.; Culbertson, C. T.; Daler, J. E.; Ramsey, J. M. Anal. Chem. 1998, 70, 3476−3480. (8) Dolnik, V.; Liu, S. R.; Jovanovich, S. Electrophoresis 2000, 21, 41− 54. (9) Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1998, 70, 3781−3789. (10) Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2000, 72, 5814−5819. (11) Molho, J. I.; Herr, A. E.; Mosier, B. P.; Santiago, J. G.; Kenny, T. W.; Brennen, R. A.; Gordon, G. B.; Mohammadi, B. Anal. Chem. 2001, 73, 1350−1360. (12) Dutta, D.; Leighton, D. T. Anal. Chem. 2002, 74, 1007−1016. (13) Griffiths, S. K.; Nilson, R. H. Anal. Chem. 2002, 74, 2960−2967. (14) Horvath, J.; Dolnik, V. Electrophoresis 2001, 22, 644−655. (15) Righetti, P. G.; Gelfi, C.; Verzola, B.; Castelletti, L. Electrophoresis 2001, 22, 603−611. (16) Doherty, E. A. S.; Meagher, R. J.; Albarghouthi, M. N.; Barron, A. E. Electrophoresis 2003, 24, 34−54. (17) Katayama, H.; Ishihama, Y.; Asakawa, N. Anal. Chem. 1998, 70, 5272−5277. (18) Hayes, M. A.; Kheterpal, I.; Ewing, A. G. Anal. Chem. 1993, 65, 2010−2013. (19) Culbertson, C. T.; Jorgenson, J. W. Anal. Chem. 1994, 66, 955− 962. (20) Chankvetadze, B.; Burjanadze, N.; Bergenthal, D.; Blaschke, G. Electrophoresis 1999, 20, 2680−2685. (21) Taylor, G. I. Proc. R. Soc. London 1953, 219A, 186−203. (22) Dutta, D.; Leighton, D. T. Anal. Chem. 2001, 73, 504−513. (23) Dutta, D.; Ramachandran, A.; Leighton, D. T. Microfluid. Nanofluid. 2006, 2, 275−290. (24) Dutta, D. Microfluid. Nanofluid. 2011, 10, 691−696. (25) Laser, D. J.; Santiago, J. G. J. Micromech. Microeng. 2004, 14, R35−R64. (26) Lazar, I. M.; Trisiripisal, P.; Sarvaiya, H. A. Anal. Chem. 2006, 78, 5513−5524. (27) Fuentes, H. V.; Woolley, A. T. Lab Chip 2007, 7, 1524−1531. (28) Toh, G. M.; Corcoran, R. C.; Dutta, D. J. Chromatogr., A 2010, 1217, 5004−5011. (29) Toh, G. M.; Yanagisawa, N.; Corcoran, R. C.; Dutta, D. Microfluid. Nanofluid. 2010, 9, 1135−1141. (30) Yanagisawa, N.; Dutta, D. Electrophoresis 2010, 31, 2080−2088. (31) Dutta, D.; Ramsey, J. M. Lab Chip 2011, 11, 3081−3088. (32) Reyes, D. R.; Iossifidis, D.; Auroux, P. A.; Manz, A. Anal. Chem. 2002, 74, 2623−2636. (33) Jacobson, S. C.; Moore, A. W.; Ramsey, J. M. Anal. Chem. 1995, 67, 2059−2063. (34) Ermakov, S. V.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2000, 72, 3512−3517. (35) Probstein, R. F. Physicochemical Hydrodynamics, 2nd ed.; John Wiley & Sons: New York, 1994. (36) Kirby, B. J.; Hasselbrink, E. F. Electrophoresis 2004, 25, 187− 202. (37) Stein, D.; Kruithof, M.; Dekker, C. Phys. Rev. Lett. 2004, 93, 035901.

reduced without compromising on its resolving power by simply decreasing the channel depth in segment D. This is because a reduction in this depth is likely to decrease the amount of band broadening arising from the nonuniform pressure-driven flow profile in the system. A smaller flow dispersion in turn should enhance the maximum resolving power of our device and also shift the location of this maximum to higher analyte velocities, allowing a reduction in the analysis time compared to that reported in this work. Fortunately, our approach to generating pressure-driven backflow in a CZE microchannel is particularly suitable for further miniaturization of the analysis column. We are currently working toward exploiting this aspect of our device in order to enhance its resolving power and expect to present results related to such a study in a future publication.

■

CONCLUSION To summarize, we have developed a novel approach to improving the resolving power of microfluidic CZE by counteracting the electroosmotic flow in its separation channel using a pressure-gradient generated on-chip. The reported device is relatively simple to design and operate as it only relies on the fabrication of a shallow region downstream of the analysis column to realize its pressure-generation capability. In our current work, we were able to improve the CZE resolution of standard dye and amino acid samples by a factor of 6−8 in a 5 μm deep analysis channel having a 0.5 μm deep shallow segment fabricated at its downstream end. It must be noted that while this shallow region was effective in suppressing fluid flow in our CZE device, it also introduced a significant amount of band broadening in the separation column. In our current design, however, the increased residence time of the analyte bands in the separation field was able to overcome the deteriorating effects of zone dispersion resulting from the pressure-driven backflow. In fact, because the latter phenomena is known to diminish with a reduction in the lateral dimensions of the separation channel, the resolving power of our CZE device is expected to improve with its miniaturization. A corollary to this statement is an anticipated deterioration in the performance of our microchip system with an increase in the depth of its analysis column. Experiments in our laboratory indeed show (data not included) that band broadening due to the pressure-driven backflow can completely offset the benefits obtained from a reduced fluid transport rate for the dye and amino acid samples used here when the depth of the separation channel in our device is increased beyond 10 μm.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research work was supported by the National Science Foundation (Grant CBET-0854179) and the Wyoming INBRE program (Grant P20RR016474).

■

REFERENCES

(1) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298− 1302. 10063

dx.doi.org/10.1021/ac302530y | Anal. Chem. 2012, 84, 10058−10063