Anal. Chem. 2000, 72, 816-820
Automated Instrumentation for Comprehensive Isotachophoresis-Capillary Zone Electrophoresis Shujun Chen and Milton L. Lee*
Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602-5700, USA
An automated comprehensive isotachophoresis-capillary zone electrophoresis (ITP-CZE) system is described. The sample is focused in the first capillary by ITP and injected repeatedly into the second smaller diameter capillary for more rapid CZE separation. Since only small portions of the concentrated zones are sequentially injected for CZE separation, overloading was not observed. Moreover, the sensitivity is enhanced because all of the concentrated zones are analyzed and the results are summed. A single detector (only for the CZE dimension) is required, and accurate timing for CZE injection is not necessary. The system was evaluated using a mixture of angiotensins. The effect of addition of leading electrolyte at the junction of the ITP and CZE capillaries before each CZE run on comprehensive ITP-CZE peak area was studied, and leading electrolyte volumes between 7 and 11 µL led to the best sensitivity. Under optimized conditions, a detection limit of approximately 5 nM could be achieved by injecting 10 µL of angiotensin solution. A major limitation of capillary zone electrophoresis (CZE) is low concentration sensitivity,1 which hinders its application to trace analysis. Different approaches have been made to solve this problem, including the use of laser-induced fluorescence or electrochemical detection and chromatographic or electrophoretic preconcentration.2 Among these, on-line electrophoretic preconcentration is the most promising. Electrophoretic preconcentration can be realized using continuous or noncontinuous buffers; however, electrophoretic concentration using a continuous buffer, also called field amplification, is only suitable for samples in nonionic matrixes.3 Isotachophoresis (ITP) preconcentration in a noncontinuous buffer is suitable for samples in ionic matrixes. The simplest configuration is to perform both ITP and CZE on the same column. With single-column ITP-CZE, the dilute sample is focused by ITP in the first section of the column, and then the concentrated sample experiences CZE separation in the remaining part of the column.4-6 In addition, a counter-flow can also be applied during * To whom correspondence should be addressed: (phone) 801-378-2135; (fax) 801-378-3667; (e-mail)
[email protected]. (1) Albin, M.; Grossman, P. D.; Moring, S. E. Anal. Chem. 1993, 65, 489A497A. (2) Chen, S.; Lee, M. L. Anal. Chem. 1998, 70, 3777-3780. (3) Chien, R.; Burgi, D. S. Anal. Chem. 1992, 64, 489A-496A. (4) Hjerte´n, S.; Elenbring, K.; Kila´r, F.; Liao, J.; Chen, A. J. C.; Siebert, C. J.; Zhu, M. J. Chromatogr. 1987, 403, 47-61. (5) Foret, F.; Szoko, E.; Karger, B. L. J. Chromatogr. 1992, 608, 3-12.
816 Analytical Chemistry, Vol. 72, No. 4, February 15, 2000
ITP focusing.7 However, this technique does not work well for trace components present in complex ionic matrixes because of CZE overloading, which destroys the CZE resolution and results in peak overlap. Coupled-column ITP-CZE may be used to overcome some of these limitations. There are two general coupled-column ITPCZE configurations: two columns coupled directly8-11 or through a T-junction.12-15 For the former configuration, either a terminating electrolyte or leading electrolyte can be used as the CZE buffer. In the latter configuration, the trailing electrolyte must be pushed away from the injection end of the CZE column before CZE is executed. Two injection methods have been used in this coupledcolumn ITP-CZE approach: pressure injection and electrokinetic injection. In both injection methods, only part of the concentrated zones is injected for CZE separation to avoid overloading. Partial injection can result in low sensitivity and poor quantitative accuracy.10,11 For ITP-CZE with a T-junction, electrokinetic injection has been used to introduce a few concentrated zones into the CZE column. To our knowledge, however, all reported coupled ITP-CZE systems can only analyze a few concentrated zones at a time to prevent CZE overloading. Moreover, two detectors are required (one for ITP and the other for CZE). The ITP detector is used to determine when the concentrated bands reach the injection end of the CZE column. Automated injection for coupled ITP-CZE systems is critical, and the use of an additional ITP detector has permitted most systems to use improved CZE injection timing routines. To analyze all of the concentrated zones in ITP-CZE from one ITP injection, we have investigated the possibility of comprehensive ITP-CZE. In comprehensive two-dimensional separations, all components separated in the first dimension are subjected to separation in the second dimension. Usually, a series of repetitive second-dimension separations are made during a single firstdimension separation, requiring that the second dimension (6) Schwer, C.; Lottspeich, F. J. Chromatogr. 1992, 623, 345-355. (7) Reinhoud, N. J.; Tjaden, U. R.; van der Greef, J. J. Chromatogr. 1993, 641, 155-162. (8) Krˇiva´nkova´, L.; Foret, F.; Bocˇek, P. J. Chromatogr. 1991, 545, 307-313. (9) Foret, F.; Sustacek, V.; Bocˇek, P. J. Microcolumn Sep. 1990, 2, 229-233. (10) Stegehuis, D. S.; Tjaden, U. R.; van der Greef, J. J. Chromatogr. 1992, 591, 341-349. (11) Reinhoud, N. J.; Tinke, A. P.; Tjaden, U. R.; Niessen, W. M. A.; van der Greef, J. J. Chromatogr. 1992, 627, 263-271. (12) Kaniansky, D.; Mara´k, J. J. Chromatogr. 1990, 498, 191-204. (13) Kaniansky, D.; Iva´nyi, F.; Onuska, F. I. Anal. Chem. 1994, 66, 1817-1824. (14) Kaniansky, D.; Zelensky´, I.; Hybenova´, A.; Onuska, F. I. Anal. Chem. 1994, 66, 4258-4264. (15) Kaniansky, D.; Mara´k, J.; Madajova´, V.; Sˇimunicˇova´, E. J. Chromatogr. 1993, 638, 137-146. 10.1021/ac990727+ CCC: $19.00
© 2000 American Chemical Society Published on Web 01/11/2000
Figure 1. Schematic of instrumentation for comprehensive ITPCZE. BR1 ) buffer reservoir 1, BR2 ) buffer reservoir 2, BR3 ) buffer reservoir 3, V1 ) valve 1, V2 ) valve 2.
separation be faster than the first. In all reported comprehensive separations, such as LC-LC,16 LC-CZE,17,18 GC-GC,19,20 and SFEGC,21,22 the two separations are performed simultaneously and injection into the second separation is from a flowing stream. However, ITP and CZE cannot be operated at the same time. In comprehensive ITP-CZE, as practiced in this study, the analytes are focused in the ITP separation and moved to the inlet of the CZE capillary. A sequential series of CZE analyses are conducted until all of the analytes have been moved from the ITP capillary into the CZE capillary for CZE analysis. A feature of comprehensive ITP-CZE that differs from other comprehensive separations is that ITP refocusing must be performed between each CZE analysis. We believe that comprehensive ITP-CZE is the first example of comprehensive electrophoretic separations on coupled columns. In this paper, technical problems for comprehensive ITP-CZE are addressed, and automated instrumentation is presented. Using this system, components in the sample are focused into zones by ITP, and the first few zones are initially sampled for CZE. Subsequent ITP focusing and CZE separation are repeated until all of the concentrated zones are analyzed. Only a single detector (for CZE) is required, and accurate timing for CZE injection is not necessary. EXPERIMENTAL SECTION Instrumentation. The instrumental setup is schematically represented in Figure 1. Two high-voltage relays with two outlets each were purchased from Kilovac (Santa Barbara, CA). High voltage could be automatically applied to any two of the three buffer reservoirs. These three buffer reservoirs were machined from Plexiglass. The screw plugs in the buffer reservoirs were made from Delrin, and the valves in buffer reservoir 1 were fabricated from Nylon. Fused silica columns of 320-µm i.d. × 430µm o.d. (14-cm total and 12.2-cm effective lengths) and 50-µm i.d. (16) Bushey, M. M.; Jorgenson, J. W. Anal. Chem. 1990, 62, 161-167. (17) Bushey, M. M.; Jorgenson, J. W. Anal. Chem. 1990, 62, 978-984. (18) Moore, A. W., Jr.; Jorgenson, J. W. Anal. Chem. 1995, 67, 3448-3455. (19) Liu, Z.; Phillips, J. B. J. Microcolumn Sep. 1989, 1, 249-256. (20) Venkatramani, C. J.; Xu, J.; Phillips, J. B. Anal. Chem. 1996, 68, 14861492. (21) Liu, Z.; Farnsworth, P. B.; Lee, M. L. J. Microcolumn Sep. 1992, 4, 199208. (22) Wu, M.; Liu, Z.; Farnsworth, P. B.; Lee, M. L. Anal. Chem. 1993, 65, 21852188.
× 190-µm o.d. (23-cm total and 16-cm effective lengths) were purchased from Polymicro Technologies (Phoenix, AZ) and used as ITP and CZE columns, respectively. Ucon-coated columns were prepared according to the procedure described elsewhere.23 A UV absorbance detector with an optical fiber assembly was purchased from Thermo Quest (San Jose, CA) for on-column detection at 215 nm. The photodiodes were purchased from Phoenix Scientific (Santa Fe, NM). The buffer reservoirs, columns, and optical fiber assembly were housed in a homemade Plexiglass box with interlock for safety. A CZE 1000R high-voltage power supply was purchased from Spellman (Hauppauge, NY). The high-voltage power supply was operated in the positive mode. The output was connected to buffer reservoir 1 or 2 through one high-voltage relay. The ground was connected to buffer reservoir 2 or 3 through the other high-voltage relay. A PHD 2000 programmable syringe pump was purchased from Harvard Apparatus (Holliston, MA), and all tubing was 1/8-in. i.d. Glass syringes (10 mL, Popper & Sons, New Hyde Park, NY) were used to fill buffer reservoirs and columns with buffer solutions. Sterile Acrodisc syringe filters (0.2µm pore size), purchased from Gel Sciences (Ann Arbor, MI), were used to remove particles and bubbles from the buffer solutions. Membrane tubing (25-mm flat width, 16-mm diameter, 12-14 000 MWCO) from The Spectrum Companies (Gardena, CA) was used to form a closed system to prevent hydrodynamic flow. A 10 µL syringe for sample injection was purchased from Hamilton (Reno, NV). The UCON-coated columns were installed in the instrument by first inserting the terminal end of the capillary through a screw plug and septum and cutting the end of the column to remove any septa residue that may have entered the column. The column was then inserted into buffer reservoir 3 such that the end of the column was visible through the Plexiglass and in contact with the electrolyte solution. The same was done for the leading end of the CZE capillary; however, the terminal end of the ITP capillary was first inserted inside, and the junction positioned in the middle of, buffer reservoir 2. The leading end of the ITP capillary was then inserted into buffer reservoir 1. The bifurcation point was connected to the counterflow via a T-channel machined in the Plexiglas of buffer reservoir 2. The channel was connected directly to the programmable syringe pump via a machined Delrin fitting which was screwed into the bottom of the reservoir and onto which the tubing from the syringe was connected. Instrument Control. A Gateway 486 PC (San Diego, CA) was fitted with an ADIO1600 card from Industrial Computer Resources (San Diego, CA). The computer was used to control the highvoltage relays and voltage output, record the current, trigger to start the syringe pump, and provide data acquisition. The card was used as follows: a 12-bit digital-to-analog converter was used to control the level of the high-voltage output, a 12-bit analog-todigital converter was used to read the current level, and a digital I/O was used to control the high-voltage relays and the syringe pump and to initiate data acquisition. Isolation amplifiers were used to isolate the control computer from high voltage. The instrument was controlled by software written in Microsoft Visual Basic. The program written in-house was run under Windows 95. An IBM-compatible PC with a Pentium microprocessor and PC1000 software purchased from Thermo Quest (San Jose, CA) (23) Ren, X.; Shen, Y.; Lee, M. L. J. Chromatogr., A 1996, 741, 115-122.
Analytical Chemistry, Vol. 72, No. 4, February 15, 2000
817
and operated under OS/2, version 3, were used for data acquisition. The data-acquisition speed was 8 Hz. Sample and Reagents. Angiotensins I, II, and III were from Sigma (St. Louis, MO) and angiotensin IV was from Bachem (Torrance, CA). Acetic acid was purchased from EM Science (Gibbstown, NJ), epoxy resin was purchased from Conap (Olean, NY), Ucon 75-H-90,000 was purchased from Alltech (Deerfield, IL), and crystal violet and dicumyl peroxide were purchased from Aldrich (Milwaukee, WI). Deionized water for the buffer solutions as well as for rinsing the reservoirs and columns was obtained from a Milli-Q water system (Millipore, Milford, MA). The leading electrolyte was 10 mM triethylamine adjusted to pH 4.5 with acetic acid. This solution was also used as the background electrolyte for CZE. The terminating electrolyte was 10 mM acetic acid. All electrolyte solutions were prepared fresh daily.24 Stock analyte solutions (160 µM) were stored in the freezer. Procedures. A flat, circular piece of membrane tubing (Spectrum) was secured at the top of buffer reservoir 3 by a screw plug and rubber O-ring, forming a closed system to prevent hydrodynamic flow. The CZE column was then filled by introducing leading electrolyte from the bottom port using a glass syringe (the same procedure is used for buffer reservoir 2). Buffer reservoir 2 and the ITP column were then filled with leading electrolyte. Finally, a 1-cm3 glass syringe with a frosted tip (Becton Dickinson, Franklin Lakes, NJ) was connected to buffer reservoir 2. The syringe was driven by the programmable syringe pump. The pump was set to “infuse” mode, and valve 1 (V1) was closed and valve 2 was opened, filling the reservoir channel with leading electrolyte. After buffer reservoirs 2 and 3 and both columns were filled with leading electrolyte, the top region of buffer reservoir 1 was filled with terminating electrolyte solution from the top. Finally, valve 1 was opened and then valve 2 was closed, thus allowing the terminating electrolyte to enter the ITP column once the sample was injected. The sample was injected using a glass syringe. The appropriate voltage was applied to buffer reservoirs 1 and 2 for ITP focusing. When the current reached a preset value where the band was approximately 1.5 cm from the injection end of the CZE column, the high-voltage relays were switched such that the voltage was turned off at buffer reservoir 2 and the appropriate voltage was then applied between buffer reservoirs 1 and 3 for CZE injection. A dye, crystal violet, was used to determine the value of the current when the band was approximately 1.5 cm away from the injection end of the CZE column. During injection of analytes from the ITP column into the CZE column, the current (measured at buffer reservoir 3) slowly decreased as analytes moved from one column to the other. After a predetermined step drop in current (∆I) was reached, the voltage was turned off at buffer reservoir 1 to suspend injection, and the programmable syringe pump was triggered to introduce an appropriate amount of CZE buffer into the junction of the ITP and CZE columns to push the remaining zones and terminating electrolyte away from the injection end of the CZE column and back into the ITP column so that CZE separation could proceed. The infusion rate of the syringe pump was 60 µL/min. Finally, the high-voltage relays were switched again and voltage was applied to buffer reservoirs 2 and 3 for CZE separation. Refocusing (24) Chen, S.; Lee, M. L. J. Microcolumn Sep. 1999, 11, 341-345.
818 Analytical Chemistry, Vol. 72, No. 4, February 15, 2000
Figure 2. Schematic representation of comprehensive ITP-CZE principles of operation. CZED, CZE detector; TE, terminating electrolyte; LE, leading electrolyte; S, sample.
of the dilute sample bands remaining in the ITP column was required before the next CZE injection because the concentrated band mixed with the buffer during CZE buffer addition from the syringe pump. These sample focusing and CZE injection steps were repeated as many times as desired. The entire instrument was automated and operated under computer control. RESULTS AND DISCUSSION The four steps during each comprehensive ITP-CZE run are schematically represented in Figure 2. The four steps are: (a) ITP focusing (or refocusing), (b) CZE injection, (c) moving the zones back away from the injection end of the CZE column by adding leading electrolyte, and (d) CZE separation. Comprehensive ITP-CZE involves repetitive runs. The first three steps involved in each run are addressed in the following sections. Constant high voltage is first applied only across the ITP column (buffer reservoirs 1 and 2) to provide initial focusing of analytes. When the current reaches a preset value, ITP focusing is stopped for CZE injection by turning off the voltage. The preset current value is chosen on the basis of previous experimental results using a dye (i.e., when the focused zone reaches approximately 1.5 cm from the CZE capillary) and is kept constant in all subsequent ITP focusing steps. Note that the current values used in this study may not be optimum for different or more complex sample matrixes. At this time, it is not known whether or not a standard set of current values can be determined and applied to a broad range of samples. A study is currently underway to examine this issue. One problem in comprehensive ITP-CZE is how to control the CZE injection amount. This problem was solved by correlating the amount injected with the change in current, ∆I, during injection. Constant high voltage is applied across the length of both columns (ITP and CZE) during CZE injection (buffer reservoirs 1 and 3). During injection, the current decreases as analytes move into the CZE column. When the current drops to a preset step value (∆I), which is determined experimentally and is input by the user, injection stops and high voltage is applied only across the CZE column (buffer reservoirs 2 and 3) for separation. Unlike the ITP focusing current preset value that is held constant for each focusing step, the injection current preset
value decreases by a value of ∆I with each injection. If the injection current was not stepped down, the next ITP band would stop just in front of the CZE column inlet without entering the CZE column. Therefore, the preset current for the next injection should be less than the current at the end of the previous injection. To gradually move the concentrated zones toward and into the injection end of the CZE column, the current must gradually decrease. The current is given by Ohm’s law
I)
V R
(1)
where I is the current, V is the applied voltage, and R is the electrical resistance of the electrolyte solution. Equation 1 indicates that the current is inversely proportional to the resistance of the solution under a constant applied voltage. The resistance of the solution at the end of the nth injection is given by
Rn )
LnT πr12κT
+
LnS πr12κs
+
LnL1 πr12κL
+
LnL2 πr22κL
(2)
where LnT, LnS, LnL1, and LnL2 are the lengths of terminating, sample, and leading zones, respectively, and κT, κS, and κL are the conductivities of terminating, sample, and leading zones, respectively. The variables r1 and r2 are the radii of the ITP and CZE columns, respectively. If the sample zones begin to move into the CZE column at the end of the n + 1th injection, the resistance is given by
Rn+1 )
L(n+1)T πr12κT
+
L(n+1)S πr12κs
+
L(n+1)S2 πr22κs
+
L(n+1)L2 πr22κL
(3)
where L(n+1)T, L(n+1)S, L(n+1)S2, and L(n+1)L2 are the lengths of terminating, sample, and leading zones, respectively. The resistance in eq 3 must be larger than in eq 2 because buffer ions and analytes of low conductivity move further into the coupled columns. In addition, eq 3 shows that the amount of sample injected can be calculated from the applied voltage and the current at the beginning and after injection. The smaller the current at the end of injection, the larger the amount of sample injected. In comprehensive ITP-CZE, the current at the end of injection is decreased by a step (∆I) after each run so that the sample zones move into the CZE column.
∆I )
V V Rn+1 Rn
(4)
The value of ∆I determines the maximum injection amount for CZE. ∆I must be small enough to prevent CZE overloading. However, the injection amount in comprehensive ITP-CZE may be different in different runs because the relationship between ∆I and the length of the ITP band injected is not linear. Despite this, multiple injections of our sample did not show columnoverloading. When leading electrolyte is used as CZE background electrolyte in traditional ITP-CZE, the CZE buffer is added at the junction of the ITP and CZE columns manually before CZE
Figure 3. Comprehensive ITP-CZE electropherograms of angiotensins using coated columns. Conditions: 14 cm × 320 µm i.d. (430 µm o.d.) and 23 cm × 50 µm i.d. (190 µm o.d.) Ucon-coated fused silica columns, 16-cm effective length for separation, 1.8-cm portion of the CZE column was inserted into the ITP column at the bifurcation point, 10-µL injection volume, 100 nM each angiotensin in terminating electrolyte, +2 kV for ITP focusing until the current reached 55 µA, then +1 kV until the current reached 21 µA, +18 kV for CZE injection, 12.0 µA current at the end of the first injection, 0.2 µA injection current step, 9-µL added leading electrolyte volume, + 19 kV for CZE separation. Peak identifications: (1) angiotensin III, (2) angiotensin I, (3) angiotensin IV, (4) angiotensin II.
separation to push the sample or terminating electrolyte solution away from the injection end of the CZE column so that CZE separation can be accomplished.9-11 However, this is impossible in comprehensive ITP-CZE because the amount of CZE buffer injected must be well-controlled. Otherwise, too much buffer volume moves the sample back out of the ITP column so that the sample is lost and the next run is impossible. This is not a problem for traditional ITP-CZE because only one run is performed. On the other hand, too little buffer volume may cause broad bands and loss of resolution. In this experiment, a programmable syringe pump controlled by computer was used for carefully adding a preset amount of buffer solution. In traditional ITP-CZE, refocusing is not necessary because, again, each sample is injected only once. However, refocusing is required in comprehensive ITP-CZE because all of the concentrated zones are analyzed, and addition of CZE buffer at the column junction causes mixing of the concentrated zones with the buffer. Refocusing of the sample is necessary before each injection. A major difficulty lies in how to estimate the position of the concentrated zones in the ITP capillary because no ITP detector is used. In this work, current measurements were used to indicate the positions of the zones. Comprehensive ITP-CZE electropherograms of a mixture of four angiotensins are shown in Figure 3. Compared with traditional ITP-CZE, more than one electropherogram was obtained for the sample. Four injections of the sample were made in order to perform a complete analysis; however, only three runs actually contained analytes. The symmetrical peak shapes indicate that overloading was not a problem for this sample matrix. According to eq 4, the smaller the current decrease, ∆I, the less sample Analytical Chemistry, Vol. 72, No. 4, February 15, 2000
819
Figure 5. Effect of added buffer volume on average ITP-CZE peak area. Conditions: same as in Figure 3 except for added buffer volume. Figure 4. Summed comprehensive ITP-CZE electropherogram. Conditions: same as in Figure 3.
injected. In practice, the minimum ∆I was limited by the noise level. In this experiment, 20 current readings were averaged to reduce the noise level, and a 0.2 µA ∆I value was used. Since the electropherograms could be summed, the sensitivity was enhanced (Figure 4). A detection limit of approximately 5 nM was achieved by injecting 10 µL of an angiotensin solution (S/N ) 4-5). It should be noted that the summation of the 4 injections using the ITP-CZE system showed some variability in retention times, as evidenced by small shoulders on peaks 1 and 4. In this study, factors affecting reproducibility were not studied since this was primarily a feasibility study of the comprehensive ITP-CZE technique. Studies on ITP-CZE reproducibility are currently underway. From our experience, in traditional column-coupled ITP-CZE, the ratio of ITP to CZE column inner diameter should be less than 4 to avoid CZE overloading. When the ratio is greater than 4, split injection must be employed. Split injection results in low sensitivity and poor quantitative accuracy.10,11 Severe overloading was not observed in comprehensive ITP-CZE even though the ratio was 6.4. This is mainly attributed to the unique injection method. Comprehensive ITP-CZE should be easily coupled to mass spectrometry because of the small CZE column diameter and the use of buffer without additives. The effect of leading electrolyte volume added after CZE injection on comprehensive ITP-CZE peak area was studied. The results shown in Figure 5 indicate that when the added leading electrolyte (LE) volume was too small (
