Anal. Chem. 2000, 72, 5774-5778
Micromachined Separation Chips with a Precolumn Reactor and End-Column Electrochemical Detector Joseph Wang,* Madhu Prakash Chatrathi, and Baomin Tian
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003
Glass microchips, integrating chemical derivatization reactions, electrophoretic separations, and amperometric detection, have been developed. The performance of the new integrated microfabricated devices is demonstrated for rapid on-chip measurements of amino acids utilizing precolumn reactions of amino acids with o-phthaldialdehyde/2-mercaptoethanol to generate electroactive derivatives that are separated electrophoretically and detected at the end-column electrochemical detector. The influence of the sample/reagent mixing ratio, reagent concentrations, driving voltage, detection potential, and other variables is explored. The integrated microsystem offers a rapid (6 min) simultaneous measurements of eight amino acids, down to ∼2.5 × 10-6 M (5 fmol) level, with linearity up to the 2 × 10-4 M level examined, and good reproducibility (RSD ) 2.2-2.7%). A step of the driving voltage is used for decreasing the migration time of lateeluting components and reducing the overall analysis time. The integrated microfabricated device expands the scope of on-chip electrochemical detection to nonelectroactive analytes and holds promise of being a powerful analytical tool. Miniaturized analytical systems, particularly “labs-on-a-chip”, have the potential to revolutionize chemical and biological analysis.1-3 Such microsystems offer many potential advantages over large analyzers, including speed, efficiency, reagent consumption, or portability. While fluorescence detection has been commonly used in microchip systems,1-4 there is considerable interest in the use of electrochemical detectors.5,6 Electrochemical detection (particularly amperometric) is an attractive choice for microchip systems because of its high sensitivity, tunable selectivity, independence of path length, and miniaturized instrumentation. The utility of on-chip amperometric detection has been demonstrated using simple separation/ detection chip layouts,5,6 but not in connection with advanced fluid manipulations (e.g., derivatization reactions), desired for full realization of complete lab-on-a-chip systems. Clearly, there is a need for enhancing the power of electrochemical detection for (1) Figeys, D.; Pinto, D. Anal. Chem. 2000, 71, 330A. (2) Jakeway, S.; de Mallo, A. J.; Russell, E. L. Fresenius J. Anal. Chem. 2000, 366, 525. (3) Haswell, S. J. Analyst 1997, 122, 1R. (4) Fister, J.; Jacobson, S.; Davis; Ramsey, J. M. Anal. Chem. 1998, 70, 431. (5) Woolley A. T.; Lao K.; Glazer A. N.; Mathies R. A. Anal. Chem. 1998, 70, 684. (6) Wang, J.; Tian, B.; Sahlin, E. Anal. Chem. 1999, 71, 5436.
5774 Analytical Chemistry, Vol. 72, No. 23, December 1, 2000
microchip separations and expanding its scope of applications. Complex solution manipulations, including on-chip precolumn and postcolumn derivatization reactions, have been documented in connection with optical detection, using a proper microfluidic network and control of the driving voltage.7,8 Such sample manipulations offer great promise for extending chip-based electrochemical detection toward important classes of nonelectroactive analytes. In the present study, we employed the reaction of amino acids with o-phthaldialdehyde (OPA)/2-mercaptoethanol for demonstrating the integrated reactor/separation/electrochemical detection microchip concept. The ability to rapidly separate and quantitate amino acids on a microchip platform is of major clinical and biotechnolgical significance. OPA derivatization has been widely used for conventional liquid chromatographic measurements of amino acids in connection with fluorescence9 or electrochemical10,11 detection. Joseph and Davies10 demonstrated that OPA/2mercaptoethanol derivatives of amino acids are electroactive and can be monitored in chromatographic effluents using modest anodic detector potentials. More recently, OPA generation of fluorescence products was applied for microchip capillary electrophoresis (CE) of amino acids.7,8 An analogous integration of on-chip derivatization reactions with CE/amperometry has not been reported. The integrated microfabricated device used in the present work had a precolumn reaction chamber, an electrophoretic separation channel, and an end-column amperometric detector. Its optimization, characterization, and attractive performance characteristics are reported in the following sections. EXPERIMENTAL SECTION Reagents. Sodium borate and dodecylsodium sulfate (SDS) were purchased from J. T. Baker and OPA, 2-mercaptoethanol (2ME), methyl alcohol, histidine, valine, isoleucine, leucine, glutamic acid, aspatic acid, arginine, and lysine were obtained from Sigma. The gold atomic absorption standard solution (1000 mg/ L) was purchased from Aldrich. All chemicals were used without further purification. The running buffer (pH 9.4) consisted of 20 mM borate buffer and 30 mM SDS. Stock solutions were prepared daily and filtered with 0.45-µm filters (Gelman Acrodisc). Sample solutions were prepared by diluting the corresponding stock (7) Jacobson, S. C.; Hergenroder, R.; Moore, A. W., Jr.; Ramsey, J. M. Anal. Chem. 1994, 66, 4127. (8) Fluri, K.; Fitzpatrick, G.; Chiem, N.; Harrison, D. J. Anal. Chem. 1996, 68, 4285. (9) Lindroth, P.; Mopper, K. Anal. Chem. 1979, 51, 1667. (10) Joseph, M. H.; Davies, P. J. Chromatogr. 1983, 277, 125. (11) Allison, L. A.; Mayer, G. S.; Shoup, R. E.Anal. Chem. 1984, 56, 1089. 10.1021/ac0006371 CCC: $19.00
© 2000 American Chemical Society Published on Web 11/01/2000
Figure 1. Schematic of the integrated reactor/separation microchip with electrochemical detection: S, R, RB, and B are the reservoirs of the sample, reagent, running buffer, and buffer solutions, respectively. See text for exact dimensions and additional details.
solutions with the electrophoresis buffer. All experiments were carried out at room temperature. The derivatization reagent was prepared by dissolving 108 mg of OPA, 76 mg of sodium borate, 80 µL of 2ME, and 2 mL of methyl alcohol in 8 mL of water. Subsequent dilutions were performed in the electrophoresis buffer as required. Chip Design. A schematic of the glass chip is shown in Figure 1, fabricated by Alberta Microelectronic Co. (AMC, model MCBF4-001, Edmonton, Canada) based on our custom design. The layout is different from that employed previously6 in that it permits precolumn derivatization reactions. The chip consisted of a reagent reservoir, a sample reservoir, a running buffer reservoir, and an injection waste reservoir. A reaction chamber (200 µm wide and 3.6 mm long) was connected through 50-µm-wide channels to the reagent and sample reservoirs at one side and to a four-way injection cross at the other side. The injection cross, was followed by a 74-mm-long, 50-µm-wide, separation channel. Apparatus. The setup of the integrated on-chip reaction/ separation/detection microsystem was similar to that described previously.6 Briefly, a Plexiglas holder was fabricated for holding the separation chip and housing the detector and the reservoirs. A short pipet tip was inserted into each of the holes on the glass chip for solution contact between the channel on the chip and corresponding reservoir on the chip holder. The amperometric detector was located in the detection reservoir (at the channel
outlet side) and consisted of a Ag/AgCl wire reference electrode, a platinum wire counter electrode, and a gold-modified screenprinted carbon working electrode. The working electrode was placed at the channel outlet; the distance between the electrode surface and the channel outlet was controlled by a plastic screw and a thin-layer spacer (50 µm). Platinum wires, inserted into the individual reservoirs, provided the electrical contact to the highvoltage power supply. This custom-made power supply had multiple voltage terminals for connection to the corresponding reservoirs. It was used for applying the selected driving voltage (between 0 and +4000 V) to a given reservoir and for switching between “separation” and “reaction/injection” modes. Amperometric Detection. The screen-printed electrodes were printed with a semiautomatic printer (model TF 100, MPM, Franklin, MA). The Acheson ink Electrodag 440B (49AB90) (Acheson Colloids, Ontario, Canada) was used for printing electrode strips. Details of the printing processes were described previously.6 A working electrode strip (0.4 in. × 1.333 in.) consisted of a carbon-ink line and its silver contact printed onto the ceramic substrate. The active working electrode area with the dimensions 0.30 × 2.50 mm was defined by a layer of insulator. The carbon working electrode area was coated with gold by applying a pulse waveform (square-wave pulse potential between -0.2 and +0.75 V, vs Ag/AgCl, with pulse width of 0.6 s) for 20 min using a solution containing 300 ppm Au(III), 0.1 M NaCl, and 1.5% HCl. Amperometric detection was performed with an electrochemical analyzer 621 (CH Instruments, Austin, TX) connected to a Pentium 166-MHz computer with 32 MB RAM. The electropherograms were recorded with a time resolution of 0.1 s while applying +0.8-V detection potential (vs Ag/AgCl wire). Sample injections were performed after stabilization of the baseline. No software filtration of the signal was used. Electrophoresis Procedure. The channels were treated before every run by rinsing with 0.1 M sodium hydroxide and deionized water for 20 and 5 min, respectively. To perform separations, the “reagent” reservoir was filled with 80 µL of the OPA/2-ME reagent solution, while the “sample” reservoir was filled with 80 µL of a mixture of amino acids. The other two ‘“buffer” reservoirs were filled with a 70-µL volume of the electrophoresis buffer. The detection/waste reservoir at the channel outlet side was filled with the electrophoresis buffer solution. A voltage of +1500 V was applied for 60 s to the reagent and sample reservoirs (with the detection reservoir grounded and other reservoirs floating), to facilitate the filling of the reaction chamber and ensure a constant mixing ratio. The sample and reagent solutions were loaded electrokinetically into the reaction chamber and mixed together by dispersion. The derivatization reaction of amino acids with OPA occurred in the reaction chamber upon mixing the reagent (in the running buffer) with the sample (in the sample reservoir), which produced the OPAamino acid derivative during the path flowing through the chamber. The reaction product “plug” was loaded into the separation channel by applying +1500 V to both sample and reagent reservoirs for 3 s with the detection reservoir grounded and the other reservoirs floating. The injection time of 3 s corresponded to an injected volume of 2 nL. Separation was usually performed Analytical Chemistry, Vol. 72, No. 23, December 1, 2000
5775
Figure 2. Electrophoregrams for a mixture containing 1.0 × 10-4 M (b) histidine, (c) valine, (d) isoleucine, and (e) leucine and 2.0 × 10-4 M (f) glutamic acid, (g) aspartic acid, (h) arginine, and (i) lysine obtained at the gold-plated screen-printed electrode. Injection potential, +1500 V; separation potential, +2000 V; injection time, 3 s; detection potential, + 0.8 V. Electrophoresis buffer, 20 mM borate buffer containing 30 mM SDS (pH 9.4). Reagent solution, 4.8 × 10-3 M OPA and 4.2 × 10-3 M 2ME. The 2ME peak is shown as (a).
by applying +2000 V to the running-buffer reservoir with the detection reservoir grounded and the other reservoirs floating. Amperometric signals of different amino acids were detected in the detection reservoir at different elution times. Different sample/reagent mixing ratios were achieved by placing different resistors (ranging from 0 to 90 MΩ) between the +1500-V power-supply terminal and the reagent reservoir (while holding the sample reservoir at +1500 V). The currents flowing through the reagent and sample reservoirs were monitored during sample/reagent loading. The actual mixing ratios of the sample and reagent solutions were calculated based on the ratios of the currents flowing through the reagent and sample arms. Safety Considerations. The high voltage power supply should be handled with extreme care to avoid electrical shock. RESULTS AND DISCUSSION The integrated reactor/separation/electrochemical microchip concept has been demonstrated for measurements of amino acids in connection with the OPA/2-mercaptoethanol precolumn derivatization and a SDS-based micellar electrokinetic separation. Figure 2 displays an electropherogram obtained at the gold-coated carbon strip detectors for a mixture containing submillimolar concentrations of (b) histidine, (c) valine, (d) isoleucine, (e) leucine, (f) glutamic acid, (g) aspartic acid, (h) arginine, and (i) lysine. (Peak a corresponds to the excess of the 2-ME reagent.) The eight amino acids peaks are well resolved, with the entire assay requiring ∼6 min; the first four amino acids are detected within less than 3 min (peaks b-e). Faster separations will be demonstrated below in connection with the use of higher or stepped driving voltages. The later is advantageous for decreasing the migration times of the late-eluting amino acid derivatives (peaks f-i). The flat baseline and low noise level indicate an effective isolation from the driving voltage. These, along with the well-defined response peaks, indicate convenient quantitation down to the micromolar level (see data below). 5776 Analytical Chemistry, Vol. 72, No. 23, December 1, 2000
Figure 3. Hydrodynamic voltammogram for 2 × 10-4 M valine (a), glutamic acid (b), and arginine (c). Reagent solution, 2.4 × 10-3 M OPA and 2.1 × 10-3 M 2ME. Other conditions, as in Figure 2.
Figure 4. Influence of separation voltage upon the response for a mixture containing 1.5 × 10-4 M histidine (b), valine (c), and isoleucine (d). Separations performed using (A) +1000, (B) +1500, (C) +2000, (D) +2500, and (E) +3000 V. Reagent solution, 2.4 × 10-3 M OPA and 2.1 × 10-3 M 2ME. Other conditions, as in Figure 2.
Optimization of the individual elements is essential as more features are integrated onto microfabricated devices. The effect of the detector potential is illustrated from the hydrodynamic voltammograms (HDVs) for valine, glutamic acid, and arginine (Figure 3). Similar profiles are observed for the three amino acids, reflecting the detection of the corresponding isoindole reaction products. The oxidation starts at +0.50 V, with a maximum response observed in the vicinity of +0.90 V. Subsequent analytical work was performed with a potential of +0.80 V that offered the most favorable signal-to-background characteristics. The different voltammetric profiles and sensitivity trend (valine > glutamic acid > arginine) of Figure 3 are related to the chemical structure of the amino acid residue of the isoindole products.10 As expected, the driving voltage has a profound effect upon the separation and detection of the individual amino acids and on the overall analysis time. Such effect is examined in Figure 4 for a mixture containing histidine, valine, and isoleucine, by increasing the voltage over the 1000-3000-V range (in 500-V increments).
Figure 5. Electropherograms recorded for sample containing eight amino acids (A) with constant separation voltage of +1500 V and (B) stepping up the separation voltage to +3000 V after 200 s of separation at a separation voltage of +1500 V. ‘Reagent’ solution, 4.8 × 10-3M OPA and 4.2 × 10-3 M 2ME. Other conditions, as in Figure 2.
Such a raise of the electrical field decreases the migration time for histidine, valine, and isoleucine from 242, 260, and 278 s to 70, 76, and 91 s, respectively. Such a decrease is in agreement with the linear dependence between the applied field strength and the migration velocity obtained for all three analytes (not shown). The peak widths (at half-height) also decrease upon increasing the driving voltage, e.g., from 6.1 s at 1000 V to ∼2.3 s at 2500 V, in the case of histidine. The decreased separation efficiency at higher fields is indicated from the decrease in the plate number (for histidine) from 9900 at 1000 V to 2900 at 3000 V. In addition to using a constant field strength, we examined a step of the driving voltage for decreasing the migration time of late-eluting components and the overall analysis time. Various methods had been proposed previously to decrease the migration time and thereby accommodate more analytes within the migration window.12,13 On-chip electrokinetically controlled gradient elution in micellar electrokinetic chromatography (MEKC)12 and open channel electrochromatography (OCEC)13 have been demonstrated. However, to our knowledge there are no reports on voltage gradient on a chip platform for decreasing the migration time. The advantage of stepping the driving voltage, over the use of a constant field strength, is demonstrated in Figure 5. By stepping the voltage from +1500 to +3000 V (after 200 s), a complete run is accomplished within 320 s, instead of 470 s under a constant field strength (i.e., reducing the overall analysis time by ∼150 s). While the voltage step results in an abrupt background noise, the baseline rapidly stabilizes to allow convenient quantitation of the late-migrating compounds. As expected, these compounds display sharper peaks following the voltage step. In contrast, the use of a constant high voltage of 3000 V resulted in largely overlapping peaks for the early-eluting compounds (e.g., Figure 4E). The derivatization reaction can be tuned by controlling the electrical fields in the sample and reagent arms and by changing the concentration of the OPA and β-mercaptoethanol reagents. (12) Kutter, J.; Jacobson, S.; Ramsey, J. M. Anal. Chem. 1997, 69, 5165. (13) Kutter, J.; Jacobson, S.; Matsubara, N.; Ramsey, J. M. Anal. Chem. 1998, 70, 3291.
Figure 6. Effect of the reagent ((A) o-phthalaldehyde and (B) 2-mercaptoethanol) concentration on the response for a sample mixture containing 1.5 × 10-4 M histidine (a), valine (b), and leucine (c). Reagent solution, 2.4 × 10-3 M OPA and 2.1 × 10-3 M 2ME. Other conditions, as in Figure 2.
Figure 7. Effect of sample/reagent mixing ratio on the response for 2.0 × 10-4 M valine (a), isoleucine (b), and leucine (c). The +1500 V voltage was applied to the sample reservoir while the voltage applied to the reagent reservoir was changed from +1390 to +1470 V (by placing different resistors between the +1500-V power supply terminal and the reagent reservoir). The mixing ratio was calculated based on measuring the ratio of currents flowing through the respective reservoirs. Reagent solution, 2.4 × 10-3 M OPA and 2.1 × 10-3 M 2ME. Other conditions, as in Figure 2.
The influence of reagent concentration was examined by varying the concentration of OPA and 2-mercaptoethanol, with a constant level of the amino acids (Figure 6, A and B, respectively). The current signals for histidine, valine, and leucine increase rapidly upon raising the OPA concentration between 0 and 1.6 × 10-3 M and nearly level off at higher reagent concentrations (A). The response increases slowly up to 5 × 10-4 M 2-mercaptoethanol, then very rapidly to a maximum value around 8 × 10-4 M, and decreases slightly at higher levels (B). The maximum reagent concentrations represent a 4-fold excess over the amino acid analytes. A large excess of OPA and 2-mercaptoethanol was used in previous chromatographic and electrophoretic applications of these reagents.7,10,11 Figure 7 displays the effect of the sample/ reagent mixing ratio upon the response for a mixture of three amino acids. For all analytes, the response rises upon increasing the sample/reagent ratio (in the reaction chamber) from 1.0 to 2.2, after which it decreases slightly. Such tuning of the mixing ratio was achieved by applying lower fields into the reagent reservoir in connection with OPA concentration of 2.4 × 10-3 M and a 2ME level of 2.1 × 10-3 M. The amperometric detector displays a well-defined concentration dependence. Such dependence was examined by recording the electropherograms for sample mixtures containing increasing Analytical Chemistry, Vol. 72, No. 23, December 1, 2000
5777
levels of histidine and isoleucine or valine and leucine in 10 steps of 2 × 10-5 M (not shown; conditions as in Figure 3 with detection potential of +0.80 V). Defined peaks, proportional to the analyte concentration, were observed for all four compounds. The resulting calibration plots were highly linear with sensitivities of 50.6, 64.1, 41.0, and 59.1 nA/mM for histidine, valine, isoleucine, and leucine, respectively (correlation coefficients, 0.992, 0.999, 0.996, and 0.995). The high sensitivity of the amperometric detector is coupled to a low noise level that results in low detection limits of 2.5 × 10-6 M valine and 2.7 × 10-6 M leucine (based on three standard deviations of the noise in assays of a mixture containing 2 × 10-5 M concentration of these compounds; conditions as in Figure 2). Such values correspond to 5.0 and 5.4 fmol of valine and leucine, respectively (i.e., 590 and 710 fg) in the 2-nL injection plugs. Slightly lower mass detection limits (e.g., 0.5 and 0.8 fmol of arginine and glycine in 0.1-nL plugs) were reported using an analogous on-chip laser-induced fluorescence detection.7 Hence, the concentration detection limits of the on-chip electrochemical and fluorescence detection modes are quite similar. Excess concentrations of OPA and 2-ME, in accordance with the optimization experiment of Figure 6, are required for attaining such micromolar detection limits (i.e., ensuring that the reagent concentration is not limiting the derivatization reaction). Good reproducibility is another attractive feature of the present integrated microfabricated device. A series of eight repetitive injections of a mixture containing 1 × 10-4 M histidine, valine, and leucine resulted in relative standard deviations of 2.3, 2.7, and
5778
Analytical Chemistry, Vol. 72, No. 23, December 1, 2000
2.2%, respectively (conditions, as in Figure 3 with detection at +0.80 V; not shown). In conclusion, we described an integrated microfabricated device that performs the derivatization, separation, and electrochemical detection processes on a chip platform. Such integration expands the scope of electrochemical detection for CE microchips to nonelectroactive analytes. The performance of the new microchip device was demonstrated for rapid and sensitive measurements of amino acids in connection with a precolumn generation of electroactive derivatives. Such an ability to rapidly separate and quantitate amino acids on a microchip platform should find important clinical and biotechnological applications. The utility of the reactor/CE/amperometry microchip for measuring other classes of compounds is under investigation in this laboratory, along with the integration of additional steps (e.g., extraction). Such a combination of CE separations, derivatization reactions, and electrochemical detection displays a potential for being a powerful analytical tool. ACKNOWLEDGMENT This project was supported by the National Institute of Health (NIH Grant RO1 RR14173-02).
Received for review June 5, 2000. Accepted October 3, 2000. AC0006371
