Anal. Chem. 1999, 71, 3828-3833
Capillary Electrophoresis and Contactless Conductivity Detection of Ions in Narrow Inner Diameter Capillaries Karl Mayrhofer, Andreas J. Zemann,* Erhard Schnell, and Gu 1 nther K. Bonn
Institute of Analytical Chemistry and Radiochemistry, Leopold-Franzens-University of Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria
Capillary electrophoresis and conductometry represent a combination of a high-resolution separation method with a sensitive detection principle for the analysis of ionic species. In this paper, results are reported that are obtained with a contactless conductivity detector. This device works without a galvanic contact of the electrolyte and the electrodes. The conductivity sensor is based on two metal tubes that act as cylindrical capacitors. These electrodes are both placed around a fused-silica capillary with a detection gap of 1 mm left in between. When a high audio or low ultrasonic oscillation frequency between 40 and 100 kHz is applied to one of the electrodes, a signal is produced as soon as an analyte zone with a different conductivity compared to the background electrolyte passes the detection gap. An amplifier and rectifier is connected to the other electrode where the signal is further processed. Limits of detection for lithium and fluoride are 4 and 13 ppb, respectively, with a linear range over 4 orders of magnitude from 90 ppb up to more than 1000 ppm for both anions and cations. Furthermore, it is demonstrated that for species with lower equivalent conductivities, such as organic ions, indirect conductivity detection is a sensitive alternative to indirect optical detection methods. Limits of detection of 50 ppb and below are obtained for organic acids. For narrow-bore analytical separation methods, such as capillary electrophoresis, detection is in most cases carried out by direct or indirect photometric absorption methods and fluorometric detection, respectively. On-column optical absorption detection is generally aggravated by the short optical path lengths, especially when capillaries with small inner diameters below 50 µm are used. Besides optical methods, mass spectrometry and electrochemical methods are the most suitable detection principles for capillary electrophoresis. Electrochemical detection techniques had already been developed in the 1970s for isotachophoretic purposes.1-3 * Corresponding author: (tel) +43-512-507-5180; (fax) +43-512-507-2965; (e-mail)
[email protected]. (1) Everaerts, F. M.; Beckers, J. L.; Verheggen, Th. P. E. M. Isotachophoresis: Theory, Instrumentation and Application; Journal of Chromatography Library Vol. 6; Elsevier: Amsterdam, 1976. (2) Everaerts, F. M. Analytical Isotachophoresis; Elsevier: Amsterdam, 1981. (3) Bocek, P.; Deml, M.; Gebauer, P.; Dolnik, V. In Analytical Isotachophoresis; Radola, B. J., Ed.; VCH Publishers: Weinheim-Basel, 1988.
3828 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
However, there are several reasons why electrochemical detection techniques, such as conductometry,4-28 potentiometry,29-35 or amperometry36-42 are less frequently used in CE compared to (4) Everaerts, F. M.; Verheggen, T. P. E. M. J. Chromatogr. 1972, 73, 193. (5) Mikkers, F. E. P.; Everaerts, F. M.; Verheggen, T. P. E. M. J. Chromatogr. 1979, 169, 11. (6) Everaerts, F. M.; Rommers, P. J. J. Chromatogr. 1974, 91, 809. (7) Gas, B.; Demjanenko, M.; Vacik, J. J. Chromatogr. 1980, 192, 253. (8) Vacik, J.; Zuska, J.; Muselasova, I. J. Chromatogr. 1985, 320, 233. (9) Foret, F.; Deml, M.; Kahle, V.; Bocek, P. Electrophoresis 1986, 7, 430. (10) Kaniansky, D.; Zelensky, I.; Hybenova, A.; Onuska, F. I. Anal. Chem. 1994, 66, 4258. (11) Mu ¨ ller, D.; Jelinek, I.; Opekar, F.; Stulik, K. Electroanalysis 1996, 8, 722. (12) Harrold, M.; Stillian, J.; Bao, L.; Rocklin, R.; Avdalovic, N. J. Chromatogr., A 1995, 717, 371. (13) Avdalovic, N.; Pohl, C. A.; Rocklin, R.; Stillian, J. Anal. Chem. 1993, 65, 1470. (14) Dasgupta, P. K.; Bao, L. Anal. Chem. 1993, 65, 1003. (15) Dasgupta, P. K.; Kar, S. Anal. Chem. 1995, 67, 3853. (16) Huang, X.; Pang, T. K. J.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1987, 59, 2747. (17) Huang, X.; Luckey, J. A.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1989, 61, 766. (18) Huang, X.; Zare, R. N.; Sloss, S.; Ewing, A. G. Anal. Chem. 1991, 63, 189. (19) Huang, X.; Zare, R. N. Anal. Chem. 1991, 63, 2193. (20) Avdalovic, N.; Pohl, C. A.; Rocklin, R. D.; Stillian, J. R. Anal. Chem. 1993, 65, 1470. (21) Alder, J. F.; Drew, P. K. P. Anal. Chim. Acta 1979, 110, 325. (22) Pal, F.; Pungor, E.; Kovats, E. Anal. Chem. 1988, 60, 2254. (23) Alder, J. F.; Drew, P. K. P. Anal. Chim. Acta 1979, 110, 325. (24) Haber, C.; Jones, W. R.; Soglia, J.; Surve, M. A.; McGlynn, M.; Caplan, A.; Reineck, J. R.; Krstanovic, C. J. Capillary Electrophor. 1996, 3, 1. (25) Haber, C.; VanSaun, J.; Jones, W. R. Anal. Chem. 1998, 70, 2261. (26) Schnell, E.; Zemann, A. J.; Volgger, D.; Bonn, G. K. Patent Pending A1016/ 97, 1997. (27) Zemann, A. J.; Schnell, E.; Volgger, D.; Bonn, G. K. Anal. Chem. 1998, 70, 563. (28) Fracassi da Silva, J. A.; Do Lago, C. Anal. Chem. 1998, 70, 4339. (29) Nann, A.; Simon, W. J. Chromatogr. 1993, 633, 207. (30) Nann, A.; Silvestri, I.; Simon, W. Anal. Chem. 1993, 65, 1662. (31) Nann, A.; Pretsch, E. J. Chromatogr., A 1994, 676, 437. (32) De Backer, B. L.; Nagels, L. J. Anal. Chem. 1996, 68, 4441. (33) Hauser, P. C.; Renner, N. D.; Hong, A. P. C. Anal. Chim. Acta 1994, 295, 181. (34) Kappes, T.; Schnierle, P.; Hauser, P. C. Anal. Chim. Acta 1997, 350, 141. (35) Kappes, T.; Hauser, P. C. Anal. Chem. 1998, 70, 2487. (36) Lu, W.; Cassidy, R. M. Anal. Chem. 1993, 65, 1649. (37) Lu, W.; Cassidy, R. M.; Baranski, S. J. Chromatogr. 1993, 640, 433. (38) Wen, J.; Cassidy, R. M. Anal. Chem. 1996, 68, 1047. (39) Gerhardt, G. C.; Cassidy, R. M.; Baranski, A. S. Anal. Chem. 1998, 70, 2167. (40) Wen, J.; Baranski, A.; Cassidy, R. M. Anal. Chem. 1998, 70, 2504. (41) Salimi-Moosavi, H.; Cassidy, R. M. Anal. Chem. 1995, 67, 1067. (42) Tenberken, B.; Ebert, P.; Hartmann, M.; Kibler, M.; Mainka, A.; Prokop, T.; Ro ¨der, A.; Ba¨chmann, K. J. Chromatogr., A 1996, 745, 209. 10.1021/ac990019o CCC: $18.00
© 1999 American Chemical Society Published on Web 07/24/1999
optical absorption methods, despite the high sensitivity for specific analytes. On one hand, the adaption of on-column optical detection methods to meet the needs of separation systems with narrowbore capillaries is easy to perform by using existing technology as long as the optical pathway is kept above certain dimensions. Furthermore, the absence of a galvanic contact of the detection electronics with the separation high voltage generally facilitates on-capillary optical detection in CE. On the other hand, capillary electrophoresis with electrochemical detection techniques, such as conductometry, potentiometry, and amperometry, requires some precautions due to the direct contact of the detection system with the separation electrolyte. In addition, miniaturization may cause disadvantages in terms of handling and equipment costs. Although these problems can be solved, they contribute to the fact that electrochemical detectors are not found in commercial CE systems, except for one end-column conductivity detector.24,25 Especially for the detection of ionic species with a high equivalent conductance, however, conductometric detection must be considered the method of choice. A capacitively coupled contactless conductivity detector for capillary electrophoresis was introduced in 1997 by our group.26,27 This detector uses either silver varnish or syringe needles as electrodes and works with a high audio or low ultrasonic oscillator frequency between 20 and 100 kHz. Recently, a similar detector also working on an oscillometric basis was reported by Fracassi da Silva and do Lago;28 however, because of its use of silver paint as electrode material, it has some disadvantages in terms of flexibility and handling. Furthermore, it was exclusively applied for the detection of cationic compounds and no efforts have been reported on the detection of anions. In the present investigation, improvements over the published versions of contactless conductivity detection in terms of detection and method development with regard to the detection of organic acids are reported. EXPERIMENTAL SECTION Instrumentation. A Quanta 4000 capillary electrophoresis system (Waters Chromatography, Milford, MA) connected to an A/D converter (UI20, Dionex Corp., Sunnyvale, CA) and a personal computer was used. Data acquisition and processing was carried out with commercially available chromatography software systems (PeakNet chromatography workstation, Dionex). To reduce baseline noise, a smoothing algorithm (Moving Average, 9 points, 10 iterations) was applied (peak height decrease and peak width increase were kept below 20%). Limits of detection (LOD) were determined at a signal-to-noise ratio (S/N) of 3:1. Fused-silica capillaries (Composite Metal Services, Worchester, U.K.) with various inner diameters (5-100 µm) and an outer diameter of 375 µm were used. Sample injection was carried out either hydrostatically at an elevation of 10 cm or electrokinetically. For the conductivity sensor, syringe needles (Hamilton Bonaduz AG, Bonaduz, Switzerland) of various lengths (0.5-2.5 cm) with an inner diameter of 410 µm and an outer diamter of 720 µm are used. Reagents. All reagents used in this investigation were of analytical grade. Standard solutions of inorganic anions (chloride, nitrate, sulfate, fluoride, phosphate, carbonate), organic anions (acetate, lactate, butyrate, formate), and cations (rubidium, ammonium, potassium, calcium, sodium, magnesium, manganese,
lithium) were prepared by dissolving the sodium salts of the anions, organic acids, and chlorides of the cations, respectively (Sigma-Aldrich-Fluka Handels-GesmbH, Vienna, Austria and Merck GmbH, Vienna, Austria) in ultrapure water (Barnstead/ Thermolyne, Dubuque, IA). Electrolyte stock solutions of 2-(Nmorpholino)ethanesulfonic acid (MES), histidine (His), sorbic acid, arginine (Arg), 18-crown-6, and chromate were prepared from analytical grade reagents (Sigma-Aldrich-Fluka Handels-GesmbH) by dissolving them in ultrapure water. For the separation of anionic compounds, 1,5-dimethyl-1,5diazaundecamethylene polymethobromide in its hydroxide form (hexadimethrine hydroxide; HDOH), which was prepared by ion exchange (Dowex-1 anion exchanger), was used as electroosmotic flow modifier in order to establish a codirectional movement of the anions with the electroosmotic flow.43-45 Electronic Parts. The detector was made from electronic components which were purchased from local electronic dealers and delivery services. The amplification unit consists of a 2-fold operational amplifier (OPA 2132, Burr Brown, Tucson, AZ). Two steps were used for amplification, two diodes (B4148) were used for rectification to direct current, and a third OP 741 allowed the regulation of the zero line. Compared to the previously published version,27 a compensation recorder can also be used for qualitative determinations. For the mains connection to the power supply, a radio frequency line filter was used in order to avoid noise from impulses and spikes. The power supply delivered a constant voltage of (15 V dc. RESULTS AND DISCUSSION It has been reported that contactless conductometry can be employed for the detection of ionic components after their capillary electrophoretic separation.27 However, the performance of this originally presented device could be further improved with regard to the limits of detection and to detector and capillary dimensions. Recent reviews on the capillary electrophoretic analysis of inorganic ions confirm that there is a demand for simple yet sensitive conductivity detection systems.46-50 Figure 1 depicts a schematic drawing of the conductivity sensor unit which is connected to the amplifier through shielded cables. The upper part schematically shows the syringe needles as well as the grounded copper foil. These parts are glued to a Perspex holder which itself is placed inside an aluminum housing to ensure good shielding as well as mechanical stability. In order to avoid a deterioration of separation efficiency, especially with short capillaries, the detection gap between the two electrodes is kept small. We found that a detection gap below 1.5 mm ensures a high separation efficiency. This can also be estimated from the zone length for sodium, which is calculated to be 1.4 cm using the electrophoretic conditions as in Figure 4. (43) Masselter, S. M.; Zemann, A. J.; Bobleter, O. Electrophoresis 1993, 14, 36. (44) Zemann, A.; Volgger, D. Anal. Chem. 1997, 69, 3243. (45) Volgger, D.; Zemann, A. J.; Bonn, G. K.; Antal, M. J., Jr. J. Chromatogr., A 1997, 758, 263. (46) Polesello, S.; Valsecchi, S. M. J. Chromatogr., A 1999, 834, 21. (47) Timerbaev, A. R.; Buchberger, W. J. Chromatogr., A 1999, 834, 117. (48) Kaniansky, D.; Masa´r, M.; Mara´k, J.; Bodor, R. J. Chromatogr., A 1999, 834, 133. (49) Masa´r, M.; Bodor, R.; Kaniansky, D. J. Chromatogr., A 1999, 834, 179. (50) Kappes, T.; Hauser, P. C. J. Chromatogr., A 1999, 834, 89.
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
3829
Figure 1. Schematic drawing of the contactless capacitively coupled conductivity detector.
To prevent a capacitive transition between the two electrodes, which would then bypass the detection gap and, as a consequence, would increase the backgound noise level, a thin piece of copper foil is placed perpendicularly between the electrodes and connected to ground. The copper foil has a small hole with a diameter of 500 µm in order to match the standard outer diameter of the fused-silica capillary (375 µm). Given an electrode length of 2 cm each, the dimensions of the conductivity sensor are approximately 5 cm × 1.5 cm × 1.5 cm. The lower part of Figure 1 demonstrates the principle of the detector with the serial arrangement of two capacitors (electrodes) and a resistor (detection gap). By means of the electrodes, a capacitive transition of current to the bore of the capillary is possible. The detection gap represents an ohmic resistance R defined by the conductance of the background electrolyte. An electric resistance is also formed by the amplification unit itself where a voltage drop is detected as soon as a solute zone with a different conductivity compared to the background electrolyte passes the detection gap. This signal is then amplified, rectified, and further processed. In the course of the present investigation, ac voltages between 10 and 11 V peak-to-peak (Vpp) were used. Various peak forms (sinus, rectangular, triangular) were generated at a preferred frequency of 100 kHz. For data processing of the detector signal, a compensation recorder, an integrator, or a chromatography data system in connection with an AD converter can be used. The capacitance of the shielded leads between detector and amplifier was approximately 50 pF which caused a sinuslike signal at the amplifier output when using a rectangular frequency of 100 kHz was used. However, it is advisable that 100 pF is not exceeded in order to avoid losses when higher oscillator frequencies are used. A direct comparison of conductivity and UV absorption detection using sorbate as a background electrolyte with a high molar absorptivity of 28 800 L mol-1 cm-1 at 256 nm51 and a low specific conductivity of 590 µS cm-1 for inorganic anions is depicted in Figure 2. It has to be emphasized that the signal-to-noise ratio of the optical detector signal for inorganic anions is even slightly inferior, despite the fact that a Hg lamp is used which produces a very intense line at 254 nm. A more fundamental comparison of photometric and electrochemical detection can be performed in terms of how the response signal is produced and recorded. Photometric absorption methods make use of the solute’s molar absorptivity as a concentrationdependent property. However, the variation of light intensity (51) Lu, B.; Westerlund, D. Electrophoresis 1996, 17, 325.
3830 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
Figure 2. Simultaneous detection of inorganic anions by direct conductivity detection (a) and indirect UV detection (b). The indirect UV signal is inversed and then amplified (10×). Conditions: capillary, i.d. 50 µm, L ) 70 cm, leff(UV) ) 62.5 cm, leff(C4D) ) 57 cm; electrolyte, 7.5 mM sorbic acid, 15 mM Arg, 0.0007% HDOH, pH 8.9; injection, 30 s at 10 cm; sample concentration, 10 ppm; separation at -20 kV; peak assignments, chloride (1), nitrate (2), sulfate (3), fluoride (4), phosphate (5), carbonate (6), acetic acid (7), lactic acid (8), and butyric acid (9).
caused by the absorption must be transformed into voltage or current first in order to be further processed. On the contrary, with electrochemical detection methods, concentration changes of an analyte directly produce a change in current (amperometry), voltage (potentiometry), or resistance (conductometry), respectively, which can be amplified and recorded without the need of an intermediate physical property. Table 1 lists the limits of detection obtained for anions at a signal-to-noise ratio of 3:1. To illustrate the sensitivity, sorbate was used as background electrolyte for the simultaneous detection of both indirect UV and contactless conductivity. When 50-µm-i.d. capillaries are used, an even lower LOD is obtained for the direct contactless conductometric detection of inorganic anions compared to indirect photometric detection using 75-µm-i.d. capillaries with sorbate as electrolytes. The direct conductivity detection response for organic acid anions, however, decreases significantly when the chain length is increased, which is due to the respective lower equivalent conductivity values. One simple way to overcome this lack of sensitivity is the use of indirect conductivity detection. As depicted in Figure 3, higher sensitivities compared to direct conductometry are obtained by using chromate as a highly conductive background electrolyte. By using this electrolyte system, it is possible
Table 1. Limits of Detection (µg/L) for Inorganic Anions and Organic Acids Comparing Indirect UV Detection, Direct Contactless Conductivity Detection, and Indirect Contactless Conductivity Detection sorbate as BGEa indirect UV
chloride nitrate sulfate fluoride phosphate formic acid acetic acid lactic acid butyric acid
chromate as BGEb
direct conductivity
indirect UV
indirect conductivity
50 µm
75 µm
50 µm
75 µm
50 µm
75 µm
50 µm
75 µm
220 351 288 119 368 214 153 209 223
86 133 114 47 212 153 91 137 160
37 79 81 44 113 108 151 564 1,110
10 22 23 13 72 44 58 217 520
418 705 681 227 1,008 473 612 692 767
88 184 168 50 218 100 148 180 206
540 352 584 65 262 116 89 99 101
156 115 283 30 103 52 39 46 42
a 7.5 mM sorbic acid, 15 mM Arg, 0.0007% HDOH, pH 8.9; U ) -20 kV; injection, 30 s at 10 cm; λ det ) 254 nm; capillary, Ltotal ) 70 cm, leff(UV) ) 62.5 cm, leff(C4D) ) 57 cm. b 2.5 mM chromate, 0.0007% HDOH, pH 8.2; U ) -20 kV; injection, 30 s at 10 cm; λdet ) 254 nm; capillary, Ltotal ) 50 cm, leff(UV) ) 42.5 cm, leff(C4D) ) 37 cm.
Table 2. Limits of Detection (µg/L) for Cations with Contactless Conductivity Detection
ammonium potassium calcium sodium magnesium lithium
50 µma
75 µma
75 µmb
18 35 31 36 24 21
5 15 11 13 10 7
4 9 8 8 7 4
a 20 mM MES/His, 1 mM 18-crown-6, pH 6.1; U ) 20 kV; injection 30 in./10 cm; capillary, Ltotal ) 60 cm, leff(C4D) ) 47 cm. b 10 mM MES/ His, 1 mM 18-crown-6, pH 6.1; U ) 20 kV; injection 30 s at 10 cm; capillary, Ltotal ) 60 cm, leff(C4D) ) 47 cm.
Figure 3. Simultaneous detection of anions by indirect conductivity detection (a) and indirect UV detection (b). Both signals are inversed and the UV signal is amplified (5×). Conditions: capillary, i.d. 50 µm, L ) 50 cm, leff(UV) ) 42.5 cm, leff(C4D) ) 37 cm; electrolyte, 2.5 mM chromate, 0.0007% HDOH, pH 8.2; injection, 10 s at 10 cm; sample concentration, 10 ppm; separation at -20 kV; peak assignments, chloride (1), sulfate (2), nitrate (3), formic acid (4), carbonate (5), acetic acid (6), lactic acid (7), and butyric acid (8).
to simultaneously perform indirect UV and indirect conductivity detection. It has to be mentioned that the detection sensitivity mainly depends on the concentration of chromate as co-ion. An optimum in terms of signal intensity is obtained using 2.5 mM chromate. With higher chromate concentrations the signal intensities decrease, whereas lower chromate concentrations cause peak shapes as well as peak resolution to deteriorate as a consequence
of the low ionic strength of the electrolyte. Table 1 also shows the respective limits of detection for inorganic and organic anions as determined by indirect UV and indirect conductivity detection. Table 2 lists the limits of detection for inorganic cations using 2-(N-morpholino)ethanesulfonic acid and histidine as the background electrolyte system. The presence of a small amount of 18-crown-6 in the background electrolyte is necessary to obtain a resolution of ammonium and potassium (Figure 4). A reduction of detection limits is eventually observed when the concentration of the buffer components is adjusted to 10 mM rather than 20 mM. One general disadvantage of photometric detection devices as well as of conductivity detectors that work with a direct galvanic contact of solutes and electrodes is the limited applicability of capillaries with very narrow inner diameters. Especially when it comes to miniaturization it would be desirable to use efficient detection techniques other than laser-induced fluorescence or mass spectrometry. However, current technology of photometric detection in commercially available instrumentation limits the practical use of capillaries with inner diameters smaller than 50 µm. One has also to bear in mind that practically the optical path length across the capillary is always less than the theoretical value due to the curved geometry. On the contrary, the inner diameter plays a less important role with contactless conductivity detection compared to other detection modes. Figure 5 shows a series of electropherograms that are obtained with capillaries of various inner diameters ranging Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
3831
Figure 4. Separation of inorganic cations with contactless conductivity detection. Conditons: capillary, i.d. 50 µm, L ) 60 cm, lC4D ) 47 cm; electrolyte, 20 mM MES, 20 mM His, 1 mM 18-crown-6, pH 6.1; injection, 10 s at 10 cm; sample concentration, 10 ppm (except magnesium and ammonium, 5 ppm); separation at +20 kV; peak assignments, rubidium (1), ammonium (2), potassium (3), calcium (4), sodium (5), magnesium (6), manganese (7), and lithium (8).
from 10 to 100 µm. To achieve equal solute plug lengths, injection was carried out electrokinetically by applying a voltage of 1 kV for 10 s. It has to be emphasized that for switching capillaries even with different inner diameters no additional modifications of the conductivity sensor itself have to be made. The capillary with the inner diameter of choice is simply led through the electrodes and the copper foil and is fixed at the required effective separation length. In order to facilitate a direct comparison of the respective electropherograms, the y-response scale was normalized to conductivity units (µS cm-1). The higher background noise of the electropherogram obtained with the 10-µm-i.d. capillary is partly due to the detector output amplification scale, which was kept constant for the various capillary diameters in this specific series of electropherograms. Furthermore, a 100-fold higher resistance is measured in a 10-µm capillary compared to a 100-µm capillary. As a consequence, the measured voltage output difference between the background electrolyte and the respective signals is smaller for narrower inner capillary diameters and has to be postamplified to a larger extent. The electrolyte solutions used throughout this investigation exhibit conductivities in the range of 400-600 µS cm-1 for direct and 200-1600 µS cm-1 for indirect conductivity detection, respectively. It is evident that the inner capillary diameters significantly affect the separation efficiencies of the investigated analytes. A reduction of the diameter from 100 3832 Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
Figure 5. Influence of different capillary inner diameters on the separation of inorganic cations. Conditions: capillary, L ) 50 cm, leff ) 37 cm; electrolyte, 20 mM MES, 20 mM histidine, 1 mM 18-crown6, pH 6.1; separation at +20 kV; injection, 10 s at 1 kV; sample concentration, 10 ppm; peak assignments, potassium (1), sodium (2), and lithium (3).
to 10 µm causes the number of theoretical plates per meter for sodium to double from 33 000 to more than 70 000. This is due to the contribution of Joule heat to zone dispersion, which becomes more marked with larger capillary diameters. Despite the fact that the measured values are not among the highest to be achieved for the separation of inorganic ionssseparation efficiencies above 1 million theoretical plates/m have been reported for inorganic ions52sone application of this detector could be the analytical investigation of small sample volumes, such as the analysis of single plant cells53,54 or single rain drops.55,56 The capacitance of the electrolyte has an influence on the peak shapes, which can be observed especially with very large signals, e.g., for the water peak as the electroosmotic flow marker. Large signals exhibit two small side signals immediately before and after the main peak that are oppositely directed to the main peak and add up to 5% of the total peak area. This effect is due not only to the lower conductivity of water but also to its high dielectric constant compared to the background electrolyte. Furthermore, various other parameters, such as the applied frequency, the wave form, the capacitive resistance of the connections between conductivity sensor and amplifier, the conductivity of the electrolyte, the length of the electrodes, and the length of the detection (52) Zemann, A. J. Chromatogr., A 1997, 787, 243. (53) Bazzanella, A.; Lochmann, H.; Tomos, A. D.; Ba¨chmann, K. J. Chromatogr., A 1998, 809, 231. (54) Bazzanella, A.; Lochmann, H.; Mainka, A.; Ba¨chmann, K. Chromatographia 1997, 45, 59. (55) Tenberken, B.; Ba¨chmann, K. J. Chromatogr., A 1996, 755, 121. (56) Tenberken, B.; Ba¨chmann, K. J. Chromatogr., A 1997, 775, 372.
gap, contribute to this effect. This effect is also less pronounced when low frequencies (e.g., 10 kHz) are used, and in this case, the reduced sensitivity of the amplifier may be compensated by an increase of the oscillator voltage. Future investigations aim at a further reduction of detection limits. However, it must be emphasized that concentration measurements in the ppt range do not only require sensitive detection devices or efficient preconcentration procedures. More-
over, cleanroom conditions or on-line sampling and analysis are strongly recommended in order to prevent sample contamination.25
Received for review January 12, 1999. Accepted May 20, 1999. AC990019O
Analytical Chemistry, Vol. 71, No. 17, September 1, 1999
3833
