Anal. Chem. 1997, 69, 2952-2954
Direct On-Line Injection in Capillary Electrophoresis Christine E. Evans
Chemistry Department, University of Michigan, Ann Arbor, Michigan 48109-1055
A direct method of sample introduction for capillary electrophoretic techniques is described using a cross configuration and high-voltage shunting. No physical disturbance of the separation capillary inlet is required, and the feasibility of direct on-line injection is demonstrated. Both full- and pinched-mode injections are evaluated, with pinched-mode injections showing superior performance. In the pinched mode, only a portion of the cross volume is introduced onto the separation capillary, as a result, a lower volume is injected, and wall effects within the cross are minimized. Preliminary studies indicate a peak height reproducibility for replicate injections of better than 4.1%, with area reproducibilities of less than 3.1% for nonoverlapping solutes. Utilizing this direct on-line injection method, many rigid or restricted capillary geometries can be accommodated, and extension to the wide range of capillary electrophoretic techniques is feasible. The reproducible introduction of sample is one of the most important aspects of any quantitative measurement technique. This injection process is even more challenging in capillary electrophoretic separation techniques, where the volume of sample must often be less than 10 nL. The two most common techniques presently used to accomplish this goal are electrokinetic injection1 and hydrodynamic injection.2 In brief, electrokinetic injection utilizes an applied electric field to move the sample into the inlet region of the capillary. In hydrodynamic injection, the sample is physically forced into the capillary inlet using siphoning, vacuum, or small pressure application. Both of these techniques are capable of producing the small injection volume required for most capillary electrophoretic separations, and their reproducibility has been extensively considered.1-9 Unfortunately, both techniques have significant disadvantages as well. The electrokinetic injection method is inherently dependent on the electrophoretic mobilities of different sample components, which can lead to a nonuniform and biased injection. Hydrodynamic injection does not suffer from (1) Jorgenson, J.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298-1302. (2) Tsuda, T.; Nomura, K.; Nakagawa, J. J. Chromatogr. 1983, 264, 385-392. (3) Rose, D. J.; Jorgenson, J. Anal. Chem. 1988, 60, 642-648. (4) Smith, R. D.; Udseth, H. R.; Loo, J. A.; Wright, B. W.; Ross, G. A. Talanta 1989, 36, 161-169. (5) Dose, E. V.; Guiochon, G. Anal. Chem. 1991, 63, 1154-1158. (6) Moring, S. E. In Capillary Electrophoresis: Theory and Practice; Grossman, P. D., Colburn, J. C., Eds.; Academic Press: San Diego, CA, 1992; Chapter 3. (7) Coufal, P.; Claessens, H. A.; Cramers, C. A. J. Liq. Chromatogr. 1993, 16, 3623-3652. (8) Fishman, H. A.; Amudi, N. M.; Lee, T. T.; Scheller, R. H.; Zare, R. N. Anal. Chem. 1994, 66, 2318-2329. (9) Shihabi, Z. K.; Hinsdale, M. E. Electrophoresis 1995, 16, 2159-2163.
2952 Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
these discriminating effects but is dependent on the precise control of the many parameters affecting Poiseuille flow and is not always feasible in packed-column separations. Another important disadvantage of these injection techniques is their limitation requiring physical disruption of the separation capillary. In both methods, the capillary must be physically moved to the sample container and then back to the buffer solution. Alternatively, the containers may be moved relative to the capillary, but in either case, physical disruption of the capillary inlet is necessary. In addition to being cumbersome, such a constraint renders some important applications intractable, including the use of different capillary geometries and the application to closed environment or high-pressure conditions. In this article, we describe the use of an on-line injection method for sample introduction that addresses many of these disadvantages. In this on-line injection method, the sample is introduced onto the separation capillary using a cross configuration and highvoltage shunting. Capillaries are arranged in a cross configuration with a small overlap volume within the cross region. Application of an electric field to the solution within one opposing set of capillary legs allows the sample to be loaded into the cross region by electrokinetic flow. High-voltage switching is then used to redirect the flow to the alternate set of opposing capillary legs, effectively injecting the sample. This injection approach is similar to methods previously introduced for etched channel separations10-13 but allows direct transfer onto an external separation capillary. Utilizing this direct on-line method, physical disruption of the separation capillary is not necessary, and many rigid or restricted capillary geometries can be accommodated. In contrast with previous capillary injection techniques, this method also provides the unique property that separations may be performed using either or both directions on the separation capillaries. This capability enhances the versatility of this injection technique, allowing simultaneous use of dual-capillary separations. Inherent in the design of previous injection techniques, only the forward direction was accessible, as the reverse direction did not exist. The on-line technique also allows samples to be manipulated without exposure to air or to outside environs, which is important for many air-sensitive or hazardous samples. Finally, this on-line injection method requires no moving parts, with the exception of high-voltage switches, rendering automation a relatively simple prospect. (10) Harrison, D. J.; Manz, A.; Fan, Z.; Lu ¨ di, H.; Widmer, H. M. Anal. Chem. 1992, 64, 1926-1932. (11) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (12) Jacobson, S. C.; Hergenro ¨der, R.; Koutny, L. B.; Warmack, R. J.; Ramsey, J. M. Anal. Chem. 1994, 66, 1107-1113. (13) von Heeren, F.; Verpoorte, E.; Manz, A.; Thormann, W. Anal. Chem. 1996, 68, 2044-2053. S0003-2700(97)00477-0 CCC: $14.00
© 1997 American Chemical Society
Figure 1. Schematic diagram of capillary electrophoresis system with on-line injector.
EXPERIMENTAL SECTION Chemicals. Standards of o-, m-, and p-nitrophenol together with mesityl oxide were obtained from Aldrich (Milwaukee, WI) and used without further purification. Sodium hydroxide and sodium borate decahydrate (Na2B4O7‚10H2O) for cleaning and buffering, respectively, were purchased from Sigma (St. Louis, MO). Preparation of buffer and standard solutions was accomplished using high-purity water that had been distilled and deionized (>18 MΩ) immediately prior to use (Model Milli-Q UV Plus, Millipore, Bedford, MA). Standard concentrations were approximately 0.2 mM for each solute. Instrumentation. The capillary electrophoresis system and injector are schematically illustrated in Figure 1. In the configuration for these studies, a microcross (LC Packings, San Francisco, CA; Model XP-75) with 75 µm i.d. fused-silica capillaries and a nominal volume of 10 nL in the cross region is connected between four reservoirs: sample, sample waste, buffer, and buffer waste. The capillary length between the cross region and the reservoirs is 30 cm for all connections. For these studies, an absorbance detector (Linear, Reno, NV; Model UVIS205) is placed on the separation capillary at 19.8 cm from the cross near the buffer waste reservoir. High-voltage connection to each of the reservoirs is accomplished using Pt electrodes. Two independent, negative-voltage power supplies (Glassman, Whitehouse Station, NJ; Model MJ30N0400) are utilized to load the sample and to drive the separation process, as shown below: reservoir sample sample waste buffer buffer waste
load mode
separation mode
ground PS1 PS2 PS2
PS2 PS2 ground PS1
Sample loading is accomplished by connecting one high-voltage power supply (PS1) between the sample and sample waste reservoirs, while maintaining the buffer and buffer waste reservoirs using a second power supply (PS2). During sample loading,
the voltage at PS2 is approximately one-half that at PS1 to maintain a near null condition at the cross region. As shown later, this voltage may also be controlled to allow a small degree of streaming flow. After the sample is within the cross region, the sample is introduced onto the separation capillary using a high-voltage switching system to change to the separation mode. This highvoltage switching system required for interchanging between load and separation modes was constructed locally. In the separation mode, PS1 is connected to the buffer waste reservoir, creating the separation potential across the separation capillary. The sample within the cross region is swept onto the separation capillary, and the separation begins. During the separation process, the sample and sample waste are controlled using PS2 at approximately one-half the PS1 voltage. In this case, the flow through the sample capillary may be maintained near the null condition. Repeated injection is simply accomplished by switching to the load mode for a brief time and then back to the separation mode. A Plexiglas enclosure surrounds this system as a safety precaution. Detector readout is accomplished using a sample integrator (Hewlett-Packard). Electrophoretic Conditions. Capillary preparation is accomplished each day by rinsing all capillaries and the cross region with 0.1 N NaOH, water, and buffer solution successively for 15 min each. A 25 mM borate solution (pH ) 9.3) is utilized as the running buffer for all measurements. Room temperature is maintained at 20 ( 0.5 °C throughout these studies, but no further attempt was made to control the capillary temperature. System configuration and high-voltage connections are as described above, with PS1 held at -10.0 kV and PS2 controlled between -5 and -4 kV. Under separation conditions, these voltage conditions result in a separation current of 25-30 µA and a current from the sample capillary between 0 and -4 µA. RESULTS AND DISCUSSION In this on-line injection scheme, sample introduction onto the separation capillary is accomplished directly and without physical perturbation of the separation capillary. As illustrated in Figure 1, the sample is loaded into the intersecting region of the microcross by application of a voltage gradient between the sample reservoir and the sample waste reservoir. Once the sample is within the cross region, the voltage conditions are switched to create a voltage gradient between the buffer and buffer waste reservoirs, implementing the separation along the separation capillary. In this way, the sample that is within the intersection region of the microcross upon voltage switching is injected onto the separation capillary. When the sample is loaded with no side flow from the separation capillary, the entire cross volume will be loaded onto the separation capillary, as illustrated below. In
this illustration, the cross-hatched lines indicate sample solution, and arrows are shown to indicate flow pathway. This full-injection Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
2953
condition is accomplished by setting PS2 to the null condition, with no current flowing through the separation capillary during sample loading. Upon switching to the separation mode, the entire volume of the cross region containing the sample is injected onto the separation capillary. As with previous etched channel studies, a small amount of streaming flow during the loading process can be used to control the sample volume injected. Illustrated below, this pinched mode of injection sections out only part of the microcross volume for transfer onto the separation capillary. This pinched condition is
Table 1. Peak Height and Area Reproducibility from Replicate Injectionsa mesityl oxide
mean SD RSD (%)
3-NP
4-NP
2-NP
area
pk ht
area
pk ht
area
pk ht
area
pk ht
1.057 1.073 1.069 1.123 1.091
5.98 6.12 6.02 5.92 6.18
2.267 2.124 2.175 2.141 2.273
5.05 4.90 4.85 4.68 4.70
1.277 1.277 1.314 1.175 1.353
2.30 2.28 2.28 2.12 2.12
2.386 2.450 2.616 2.632 3.122
3.85 3.70 3.75 3.65 3.60
1.089 0.025 2.3
6.04 0.11 1.7
2.178 0.067 3.1
4.84 0.15 3.1
1.280 2.220 2.705 3.710 0.076 0.092 0.29 0.096 6.0 4.1 10.7 2.6
a PS1 ) -10.0 kV (i sep ) 28 µA); PS2 ) -4.71 kV (iinj ) -1.5 µA); injection time ) 30 s. All values expressed in arbitrary units.
accomplished by setting PS2 to yield a small negative current, creating a small flow in both legs of the separation capillary toward the microcross during the loading cycle. Thus, the pinched mode describes not the voltage condition but the side flow from the buffer and buffer waste during the load cycle, as measured by the current. The current through the sample capillary is significantly greater than that within the separation capillary, yielding an overall flow trajectory from the sample to the sample waste reservoirs. This results in a “pinching” of the sample within the cross region. Switching to the separation mode transfers this diminished volume of sample onto the separation capillary. Consistent with previous studies in etched channels, preliminary results indicate a time-dependent injection volume for the fullinjection condition, presumably due to sample diffusion into the separation capillary. Also in agreement, time-independent behavior is observed with pinched conditions once the initial streaming conditions are in place. This time-independent behavior, together with the ability to decrease detrimental wall effects within the cross region, creates a significant advantage for pinched-mode injections. Representative peak height and peak area reproducibility data for replicate on-line injections are shown in Table 1 for the separation of nitrophenol isomers using mesityl oxide as the void marker. Under moderate focusing conditions (PS2, iinj ) -1.5 µA), both the area and peak height reproducibility are excellent for the void marker and for 3-nitrophenolate. Although the area reproducibilities for the 2- and 4-nitrophenolate are not as favorable due to overlap between these two solutes, the peak height reproducibility is quite good for both solutes. In each case, relative standard deviation values are in good agreement with the ∼2% RSD achieved by hydrodynamic and electrokinetic injection methods.1-9 Moreover, replicate injections were accomplished using only high-voltage switching, with no physical disturbance of the separation capillary. (14) Evans, C. E. J. Microcolumn Sep., submitted. (15) Paladini, A. A.; Silva, J. L.; Weber, G. Anal. Biochem. 1987, 161, 358-364. (16) Masson, P.; Reybaud, J. Electrophoresis 1988, 9, 157-161. (17) Masson, P.; Areiero, D. M.; Hooper, A. B.; Balny, C. Electrophoresis 1990, 11, 128-133. (18) Somero, G. N. Annu. Rev. Physiol. 1992, 54, 557-577. (19) Yonker, C. R.; Smith, R. D. J. Chromatogr. 1990, 517, 573-81.
2954 Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
Successful implementation of this injection method in a closedsystem configuration has been recently demonstrated.14 Not presently feasible for capillary separations using other injection techniques, a closed system may be desirable when samples or reagents present a significant health hazard to the operator or for those samples that may be air-sensitive. Moreover, utilizing this on-line injection method in a closed configuration, the feasibility of high-pressure capillary electrophoresis is now being investigated. Previously operated only under atmospheric conditions, electrophoretic measurements in a high-pressure environment have many advantages for fundamental as well as practical studies.15-19 CONCLUSIONS A method for the direct on-line injection of samples is successfully demonstrated for electrophoretic applications in a capillary geometry. Using a cross configuration and high-voltage switching, the sample is introduced without physical disruption of the separation capillary inlet. Peak height and area reproducibilities are achieved that are consistent with values previously obtained for electrokinetic and hydrodynamic methods. This online injection technique offers many advantages over these more commonly utilized injection methods for capillary separations. Since no physical disruption of the separation capillary inlet is required, injection in previously intractable rigid or restricted capillary environments is now feasible. Moreover, on-line injection makes possible the operation of capillary electrophoretic separations in the closed-system configuration necessary for air-sensitive or very hazardous samples. Although described here in a singleinjection mode, this on-line injection technique may be readily implemented in a series geometry. Such multiplexed injection options show considerable promise for application to those simultaneous parallel separations of advantage for combinatorial approaches to analysis. ACKNOWLEDGMENT Partial support of this research was provided by The Eli Lilly Co. and is gratefully acknowledged. Received for review May 8, 1997. 1997.X
Accepted May 21,
AC9704772 X
Abstract published in Advance ACS Abstracts, July 1, 1997.
