Capillary electrophoresis

1.2. Capillary format separations

It is obvious that there is a strong trend in analytical chemistry toward developments that address research problems in biomedical areas.  This has resulted in the development of tools that determine levels of nonvolatile analytes such as polar drugs, proteins, DNA, and biomarkers.  These developments have been partially propelled by capillary format separations such as capillary zone electrophoresis (CZE), capillary electrochromatography (CEC), and the ability of mass spectrometry to ionize large biomolecules via electrospray ionization.

Capillary format separations have an obvious advantage for pollution prevention (low volume of solvent use for separations) and therefore provide a Agreen@ chemistry approach to analysis.  They also exhibit high mass sensitivity (amount of sample on-column that can be detected), but a concomitant increase in concentration detection limits due in one sense to the path length limitations described and the practical limits to the volume of sample that may be injected.  This relatively high concentration detection limit is a primary concern for environmental analysis where normally large amounts and volumes of samples are available.  Typical detection limits ranging from mg/kg to ng/kg are often realized in commonly used techniques such as GC/FID, GC/ECD, GC/MS, and HPLC/DAD or HPLC/FLD, either directly from a solution of the sample or sample extract or by using preconcentration.  A useful figure for target analyte amount for sample injection is 1 pg / μL that is currently practical for GC/MS methods and 1 fg/nL for CE methods.  Capillary format separations are, in a practical sense, limited by the volume of injection possible, but this deficiency is offset by the advantage of a plug-like flow profile which results in sharper peaks and greater selectivity [9] from the rather limited practical range of ion mobilities .  In the case of optical methods of detection, the optical path length is also limited by the geometry of the capillary in any practical detection scheme.  Recent work in addressing the short optical path length has been quite creative.  A Az-cell@ introduces a straight detour into the inside of the capillary, allowing the laser to interact with the contents along a section of the capillary length rather than through a cross-section. [10] 

LIF has assumed an important role as a detector for CZE and other capillary techniques because of its inherent sensitivity and resultant low detection limits (typically, 10-7 M to 10-13 M) [11].  Ultraviolet detection methods are unable to achieve these low detection limits, typically being limited to about 10-6M - 10-7M.  For organic ions, the power of LIF increases our ability to sensitively screen for nonvolatile and other analytes.

1.3. Derivatization of target analytes with fluorophores

Since there is a requirement for the target analyte to exhibit fluorescence, some analysts have resorted to derivatization with fluorophores to provide the desired properties.  In principle, this approach would produce a derivatized analyte capable of being detected at very low levels.

Two limiting considerations must be addressed within this context, however.  One concerns the lowest practical concentration of analyte that will undergo reaction with the derivatizing reagent.  Chemical kinetics dictates that some analytes will not react appreciably at concentrations of interest, but must first be preconcentrated in order for the reactions to proceed at useful rates.

The second consideration must address the production of artifacts, coupling products, and derivatives from coextractives, all of which put additional burden on the separation system.  To address this very challenging problem two approaches may be pursued.  First, one may apply cleanups to limit the number of coextractives present.  Second, one may develop cleanups to remove the bulk of the unwanted side products and other fluorescing background from the reaction mixture.

1.4. Quality Assurance/Quality Control (QA/QC) aspects of CE/LIF

Two primary questions arise in the context of QA/QC issues:  How reproducible are the run to run migration times of target analytes?  How transferable and robust are methods developed in one laboratory when ported to another laboratory (or collaborative study)?  Unfortunately, the performance to date for both CZE and micellar conditions has not been as good as those methods based on GC/MS, HPLC, and TLC.

The well known variation in electroosmotic (EO) flow has been a major source of irreproducibility.   One QA/QC tool that has been used is to employ an internal standard as both a quantitative tool and as a corrective tool for migration time variations, especially when those variations are a result of EO flow variations.  Migration time (MT) corrections can be applied based on the reciprocal relation between intrinsic mobility and MTs [12].  Thus, if a typical standard run is used as a benchmark for the MTs of internal standard and analytes, subsequent runs of samples can be corrected back to this benchmark where the correction factor is based on the relationship between the MT of the internal standard in the two runs.  This factor can then be used to correct the analyte MTs to what they would be if the EO flow was the same as the benchmark run.  Typically, corrected MTs of analytes can be as good as 0.3 % reproducibility.

A more difficult problem to deal with is the change in selectivity that is observed when methods are transferred to a different laboratory.  Separations can no longer be maintained, or the order of analytes that reach the detector is altered.  The reproducibility issue is particularly acute under micellar conditions.  Adsorption problems also occur in free zone, particularly when non-borate buffers are used or when dealing with ions that are subject to these problems.

Corrected peak areas must be used when on-column detection is implemented.  This is usually automatically handled by the data system software, but the analyst may have to factor this in for laboratory built systems.  This is a direct result of the difference in time that the analyte spends within the detector window as a function of its apparent mobility as it moves through the capillary.

It is considered good practice to bracket the sample series of runs with standards, validating performance and adherence to the calibration plot.  Analysis of reagent blanks (taken through the method) to demonstrate the lack of contamination and solvent blanks between samples and standards is also recommended.

1.5. Environmental analysis

A number of examples will illustrate the attributes and advantages of CE/LIF for solving the analytical problems posed by various environmental analysis scenarios.  The performance or potential of CE/LIF will be compared with the currently used methodology.  

 1.5.1. Dye tracers in groundwater migration studies

Groundwater migration using fluorescent dyes presents an analytical problem almost ideally matched to CZE/LIF.  The anionic dyes migrate in such a manner that they follow the cations and neutrals in entering the detection window (longer MTs).  CZE/LIF has been used for tinopal (near UV, 354 nm excitation) and fluorescein (visible, 488 nm excitation) dyes in actual groundwater migration [13, 14, 15].  These applications may result in improved detection limits, specificity, and quality control.  Multi-wavelength lasers may be applied for multiple dye injection studies [16].  Typical dyes used include fluorescein, tinopal (fluorescent brightener), eosin, and rhodamines.  Fig. 2 illustrates results for a groundwater migration study using fluorescein as the indicating dye.

1.5.2. Solid waste and contaminants in solid matrices

It is possible to screen for a wide range of analytes that fluoresce using frequency-doubled lasers operating in the deep UV (e.g., 257 nm).  Alkyl phenols and hydroxy-PNAs can be detected as migrating anions and therefore separated from neutral hydrophobics [17].

Conversely, organic bases can be screened by using acidic conditions for their separation and detection as cations.  These separations are particularly simple, but may be complicated by surface adsorption and coelutions.

Micellar agents or cyclodextrins offer alternatives and are needed for the resolution of neutral molecules using a technique called micellar electrokinetic chromatography (MECC).  This is accomplished for PNAs based on separations using cyclodextrins with LIF detection resulting from the HeCd laser operated at 325 nm [18].

Another variant employs capillary electrochromatography, a hybrid of capillary HPLC and electrophoresis wherein the mobile phase flow is generated electrically as the electroosmotic flow [19].

1.5.3. Atmospheric contaminants

Capillary electrophoresis has been applied to various substances of interest to atmospheric chemistry and particulate matter [20].  Recent work by Dabek-Zlotorzynska et al. shows that CE/LIF can be used to measure dimethylamine and other low-molecular-weight amines in atmospheric aerosol studies [21].

1.5.4. Chemical characterization of matrices

The characterization of a sample can be limited to a list of analytes of interest or it can be expansive enough to encompass the character of the matrix itself.  Humic substances include fulvic acid, humic acid, and humin.  These constituents, while not frequently considered in environmental analyses, play an important role in the way contamination binds to or flows through soil matrices.  CE has been used to study the interaction of free metal cations with fulvic acid [22].  It has also been used to characterize sewage effluent for fluorescent acids [23].

1.5.5. Drinking water and groundwater contaminants

An area that has stimulated considerable recent interest both publicly and analytically is that of endocrine disrupting compounds (EDCs).   There is a demand for reliable analytical methods capable of detecting trace levels of many pesticides, polychlorinated biphenyls, and dioxin-like compounds, among others.   CE/LIF may play a role in applicable cases providing some type of fluorescent property belongs to the target or can be added to the target analyte.  This focus on EDCs has resulted in closer scrutiny of many pesticides and other suspect compounds such as polychlorinated biphenyls.  The Safe Drinking Water Act (SDWA) and the Food Quality Protection Act (FQPA) have both adopted new regulations for monitoring thousands of compounds that interfere with the human and ecological hormone systems.

An example of the application of CE to pesticide analysis is the use of  MECC with LIF to analyze for trace levels of phenoxy acid herbicides [24].  This highly sensitive method requires a complex derivatization but is able to achieve femtomole detection levels.  CE even has the power and sensitivity to separate enantiomers of phenoxy acid herbicides [25]. CE/LIF in an immunoassay format was also applied to 2, 4-dichlorophenoxy acetic [26].  CE/LIF has been used to determine anilines [27] and aliphatic amines [28].  Surfactants have also been determined using CE/LIF operating in the UV region [29, 30, 31].

1.5.6. Biomarkers of exposure

Biomarkers of exposure can be monitored by analyzing certain physical response parameters such as protein adducts, DNA adducts, and other biological indicators.  This application seeks to exploit the sensitivity and specificity of LIF and the separation power of CE.  However, the area has yet to undergo significant development.

1.5.7. Food contaminants

The complexity of food analysis is a combination of difficult matrices and low detection level requirements.  In addition, the contaminants present in food may be pesticide residues, natural toxins, such as mycotoxins, or the residues or breakdown products of food additives.  CE with fluorescence detection has been used to determine the levels of fumonisin B1 [32].  Certain food contamination issues have become matters of widespread interest to the media and the public.  The low detection limits achieved by CE/LIF make it an excellent choice for food analysis and monitoring where applicable.

MECC with LIF has been used in a study of aflatoxin contamination in corn.  In a laboratory study, aflatoxin spores were introduced onto corn kernels and allowed to grow for two weeks at room temperature.  The complex sample matrix yielded chromatographic peaks that were difficult to resolve.  MECC data showed separation times of less than a minute and, at the time of the study, the detection was the limiting factor [33].

1.5.8. Emerging developments

   1. Multidimensional separation:  Multidimensional separations allow more complete analyses because analytes are separated by more than one method, e.g., GC and (LC), supercritical fluid extraction (SFE) and LC, or other combinations.  Though in theory, less efficient separation is obtained with two-dimensional chromatography, the advantage in separation power greatly outweighs any loss in efficiency [34].  The resolution obtained depends on several factors: the orthogonality of the methods, the effectiveness of transferring from one column to another, and the completeness of the whole sample dispersion.

CZE has been used with reversed-phase HPLC in an automated comprehensive two-dimensional method.  Each separation phase is capable of effecting separations using a different separating principal: they are orthogonal methods.  In the first report of an electromigration injection from a flowing stream, Bushey and Jorgenson [35] presented the advantages (excellent separation, ease of injection into the CZE system) and disadvantages (long analysis time, inefficiency in sampling from the first column).

CEC has been used to separate 16 different polycyclic aromatic compounds (PAHs).  In CEC, an electric field is applied across columns that are packed with microparticulates.  The electroosmotic flow becomes a tool for chromatographic separations [19].  CEC has an advantage over CZE in this application because it is capable of separating many uncharged species.  However, CEC has yet to be significantly adopted in environmental analysis and seems to suffer from lack of robustness, undue complexity, expensive columns, and poor reproducibility.

2. Materials

2.1. Chemicals   

All organic compounds were obtained from Aldrich Chemical Company, Inc. (Milwaukee, WI, USA) and Molecular Probes (Eugene, OR) unless otherwise specified.  Other chemicals were from standard sources of supply, and all were used as received.  Deionized water (18 mohm quality) was used for all aqueous solutions.  Buffer solutions were freshly prepared at least weekly.  Solutions of dye standards were prepared from solid dye and serially diluted.

2.2. Optics and fused silica

Fused silica glass with polyimide coating was obtained from PolyMicro Technology, Phoenix, AZ, USA.  Capillaries of 0.050 mm ID and 0.075 mm ID are acceptable for separations.  Optics obtained commercially should be appropriate for the application.  If wavelengths below 365 um are used, then fused silica optics should be employed with appropriate coatings for the application wavelength region. 

3. Methods

3.1. Capillary electrophoresis of fluorescein dyes

A P/ACE Model 5000 Capillary Electrophoresis System (Beckman Instruments, Fullerton, CA, USA) was used for all commercial instrument electrophoretic determinations reported here.  The instrument was fitted with a capillary 57 cm in total length (50 cm from the origin to the detector window) and 75 μm inside diameter.  Detection was accomplished with an Ar ion laser operated at 488 nm emission and detection using a notch filter (488 nm) and a band pass filter (520DF20). Unless otherwise noted, electrophoresis was carried out at 30 kV.   The temperature of the capillary was maintained at 25EC.  The capillary was equilibrated with running buffer for two minutes prior to beginning of an experiment, and washed for two minutes with alkali and water between runs.  Migration times, peak widths, and peak areas were determined directly from peaks displayed by the data system or by processing software.  Corrected peak areas, as computed by the instrumental software were further normalized by dividing them by the area of the peak of the internal standard, which was erythrosin B.  This computation was necessary to correct for the small variations in injection volumes resulting from the pressure injections typically of 5 sec duration (nominally 20 nL).

3.2. Sample handling 

Samples containing dyes or other fluorophores can be preconcentrated from the matrix by several procedures.

3.2.1. Calibration Curve

A regression analysis was carried out on the ratio of corrected areas (fluorescein corrected area divided by internal standard corrected area) versus the ratio of fluorescein concentration to internal standard concentration from 1 X 10-7 M to 1 X 10-10 M in half decade increments resulting in a 7-point calibration curve.  Both unforced and forced through the origin regressions were considered and both resulted in correlation coefficients of 0.99.  The equation of the line was used to calculate the concentrations of fluorescein in the samples and check standards run during the course of analysis.  The concentration of fluorescein was calculable based on the known volume and concentration of internal standard added to the sample and the known volume/weight of the sample being analyzed.

3.2.2. Quality Assurance/Quality Control (QA/QC)

Each group of samples to be analyzed was bracketed before and after by a representative standard/internal standard QC sample to establish adherence to the calibration curve equation and MT variations due to changes in EO flow and therefore MT.  Deviations greater than 15% in agreement with the calibration curve results would be cause for rerunning of standards, construction of a new calibration curve, or replacement of the capillary.

The migration time variation on a given day/capillary was approximately 5% in agreement with the standards of that day.  The variation provides a rough window for anticipating the response of the internal standard and fluorescein.  Using MT correction based on the internal standard, the expected and measured MT of the fluorescein peak fell within 0.3%.  The position of both the internal standard and fluorescein could also be estimated from the position of the EO flow disturbance.  The EO flow could be seen as a peak on the optical scale or can be monitored as a change in current as the injection plug exits the column.  Further confirmation of the internal standard peak can be obtained by overspiking the sample with an additional aliquot of internal standard and observing the appropriate increase in peak area.  Fluorescein itself could be confirmed by a similar procedure.  In the samples reported in this work, no problems were encountered in identifying internal standard and analyte responses due to the low background level observed with CZE/LIF.

3.2.3. Optical bench experiments with optical brighteners

 (An optical bench with components and light-tight enclosure was constructed and used for all LIF experiments.  The overall design was based on that of Nie et al., [11] but modified in certain respects.  A special cylindrically symmetric capillary holder was machined to insure optical alignment with the laser beam, lenses, slit, and the newly installed capillary with window made by removing the polyimide coating.  The entrance lens was an L-50X and the exit lens was an M-60X (Newport, Irvine, CA).  A mass spectrometer slit (VG 7070EQ source slit, Micromass, Beverly, MA) was mounted in a special holder to exclude wall fluorescence transmitted to the detector as previously described. [3]  The bench was fitted with a fused-silica capillary Polymicro Technologies, Phoenix, AZ) 57 cm X 75-m I.D., 50 cm to the detector, with LIF detection using the 354-nm line of the HeCd ion laser, 3 mW, model 7203N (Liconix, Santa Clara, CA ) and two 450DF100 (i.e., 50% transmission at 400 nm and 500 nm) emission filters (Omega Optical, Brattleboro, VT) in series.  The power supply for CE was a Series EL (Glassman, White House Station, NJ).  The temperature of the capillary was 25°C, and electrophoretic runs were about 10 minutes at 20 kV using a 40 mM borate buffer at pH 9.1.  The buffer was prepared by weighing 0.381 g of sodium tetraborate decahydrate followed by dissolution in 100 mL of deionized (DI) water.  The capillary was equilibrated with running buffer at the start of each experiment, and washed extensively (minimum 2 min each) with 0.1 M sodium hydroxide, DI water, and running buffer between analyses.  Rinsing was accomplished by using capillary rinse reservoirs (SGE, Austin, TX) at about 20 psi of nitrogen pressure by fitting the injection end of the capillary through septum seals on the reservoirs.  Migration times, peak widths, and detection limits were either read directly from the monitor or from printouts of the data system (Austin P-90 computer, Austin, TX loaded with Beckman System Gold, Ver. 8.1, Fullerton, CA) with data acquisition using a Beckman 406 analog interface (2 V full scale output).  The photomultiplier tube (PMT) was model R928 (185-900 nm) fitted with socket E0719-21 (Hamamatsu Photonics Systems, Bridgewater, NJ) and was operated at 900 V with power supply model 230-03R (Bertan, supplied by Hamamatsu).  The current amplifier for the optical signal was a model 428 (Keithley Instruments, Cleveland, OH) and used the auto-current suppression facility of the amplifier to zero the background signal and maintain full amplifier dynamic range (0 to 10 V output).  Corrected peak areas, as computed by using a spreadsheet (peak area multiplied by the velocity of the ion [length to the detector divided by time]), were normalized to the corrected peak area of the internal standard (7-hydroxycoumarin-4-acetic acid) as a control for the variations in the nominal volumes of the gravity injections (10 sec to 40 sec at 30 cm height corresponding to about 40 to 170 nL). A microampere electrometer with 0 to 1 V output (1 V = 200 μA) was constructed for measuring current through the capillary and was also interfaced to the Beckman 406 ADC to provide a record of the electrophoretic current.

Four different dyes (Tinopal CBS-X, fluorescein (acid yellow 73), rhodamine WT, and eosin Y) were injected into four wells at a RCRA site and were monitored at three wells at a nearby Superfund site.  Each dye (10 - 30 lbs) was injected with 2000 L of water resulting in a 10 mM concentration level for each dye.  Thereafter, 8000 L of water was used to flush the dyes into the surrounding groundwater.  Samples were taken before injection and for about two months afterward resulting in about 22 samples.  Samples consisted of vial samples of water, "receptors", and 1-L water samples at the monitoring wells.  The "receptors" consisted of fiberglass mesh filled with coconut charcoal and weighted to remain near the bottom of the well.  The standard protocol called for 1 g of charcoal from the receptor to be extracted with 10 mL of a solution consisting of 5:3:2 (propanol: water: concentrated ammonium hydroxide).  Results for fluorescein may be reported as ppt-levels in the 10-mL extractant of the pads or ppt in the water when determined directly from a portion of the water sample.

3.2.4. SPE Sample Handling        

Fluorescein was isolated from spiked DI water samples or groundwater samples using SPE with styrene-divinylbenzene (SDVB) extraction disks.  The disks were prepared following the manufacturer’s directions by soaking in 10 mL acetone and then pulling the solvent through the disk.  The process was repeated with 10 mL methanol and then water without letting the disk become dry.  Samples were then added, adjusted to pH 5.0, and pulled through at 25 mm Hg vacuum.  The disks were dried for 2 min and then eluted twice with 6 mL of methanol.  The methanol eluant was concentrated as necessary with a gentle stream of nitrogen with gentle warming to achieve a recovered concentration within the detection limits of the CE/LIF technique.

4. Notes

1. New capillaries should be conditioned with generous rinsing with 0.1 N NaOH solution and DI water.  If the capillary does not perform as expected, it should be discarded and a new one installed in either the commercial holder or the user designed apparatus.  Although commercial capillaries are available in a variety of coatings, their expense compared to the bare silica ones is considerable.  The use of inexpensive columns of bare fused silica is one of the strengths of the approach. Unstable current at the operating voltage can be due to a poor capillary as well.  As long as the Joule heating (V times i) does not exceed the heat dissipation available to the setup (2.5 W or lower for air-cooled setups), a relatively stable current should be readily achieved.  Slow drift of current up or down is not usually a problem for concern.  Note also that current will dip and then change after the passage of the EO flow disturbance through and out of the capillary.
2. An excellent buffer for CE is borate either as the titrated pH 8.3 (tetraborate versus boric acid) or the tetraborate itself (pH 9.2).  Many other buffers are used but often with poorer performance and ruggedness compared to borate.
3. Remember to replace the anode and cathode buffers on a regular basis because of depletion of ions, sample elution from either intrinsic mobility or EO flow, and electrical processes.  Continuity must be maintained so a buffer-filled capillary is required and the introduction of a plug of nonconducting liquid or easily boiled liquid could lead to no conductivity.  Remember to rinse to a waste container, not the receiving buffer.  For pressure injections do not exceed about 10 % of the capillary length (usually less than 20 sec, typically 5 sec).  Use a diluted buffer for the sample about 10% of the ionic strength of the working buffer for focusing effects.  Running buffer strengths of 10 to 50 mM are often sufficient for good results.  Avoid working in the pH 5-7 region if possible because of the strong dependence of EO flow in this region on pH (i.e., the extent of ionization of the silica wall is changing steeply in this region as a function of pH). 
4. Doubling the length of the capillary will result in MTs four times longer if all other parameters are kept constant. Efficiency is proportional to field strength if Joule heating is controlled.
5.  Migration time variations are often a consequence of EO flow variations from run to run and capillary to capillary.  Use internal standard for MT corrections arising from EO flow variations and for quantitation.  This also serves QA/QC purposes for assessing the performance of each run including MT and relative response.  Use corrected areas to account for on-column detection bias (simply area / MT).

Notice

The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD), funded the work involved in preparing this article.  It has been subject to the Agency=s peer review and has been approved for publication.  The U.S. Government has the right to retain a non-exclusive, royalty-free license in and to any copyright covering this article.

References

1. Daughton, C. G., Ternes, T. A., (1999) Pharmaceuticals and personal care products in the environment: Agents of subtle change? Environ. Health Perspectives 107(suppl 6), 907.
2.  http://www.epa.gov/ppcp/
3.  http://www.epa.gov/nerlesd1/chemistry/org-anal/home.htm
4. Brumley, W. C. and Winnik, W. (1996) Applications of Capillary Electrophoresis/Mass Spectrometry to Environmental Analysis. in Applications of Liquid Chromatography/Mass Spectrometry in Environmental Chemistry. D. Barceló, ed., Elsevier, Amsterdam, C12, p481-527.
5. Heiger, D. N., (1992) High Performance Capillary Electrophoresis, Hewlett-Packard 12-5091-6199E, Walbronn, Germany.
6. Brumley, W. C. (1995) Techniques for handling environmental samples with potential for capillary electrophoresis. J. Chromatogr. Sci. 33, 670-685.
7. Gooier, C., Kok, S.J., and Ariese, F. (2000) Capillary electrophoresis with laser-induced fluorescence detection for natively fluorescent analytes Analusis, 28, 679-685.
8. Chervet, J. P., van Ling, R., Evans, K., and Salzmann, (1999) J.-P. Capillary and nano HPLC using a dedicated instrument. Am. Lab. August, 44-50.
9. Dadoo, R., Yan, C., Zare, R.N., Anex, D.S., Rakestraw, D.J., and Hux, G.A. (1997) Advances toward the routine use of capillary electrochromatography. LC-GC, 15(7) 630-635.
10. Chervet, J. P., van Soest, R. E. J., and Ursem, M. (1991) Z-Shaped flow cell for UV detection in capillary electrophoresis. J.  Chromatogr. 543, 439-449.
11. Nie, S., Dadoo, R. and Zare, R. N. (1993) Ultrasensitive fluorescence detection of polycyclic aromatic hydrocarbons in capillary electrophoresis. Anal. Chem. 65, 3571.
12. Brumley, W. C. and Brownrigg, C. M. (1993) Electrophoretic behavior of aromatic-containing organic acids and the determination of selected compounds in water and soil by capillary electrophoresis. J. Chromatogr. 646, 377-389.
13. Brumley, W. C., Ferguson, P. L., Grange, A. H., Donnelly, J. L., and Farley, J. W. (1996) Applications of capillary electrophoresis/laser-induced fluorescence detection to groundwater migration studies. J. Cap. Electrophoresis 3, 295-299.
14. Ferguson, P. L., Grange, A. H., Brumley, W. C.,  Donnelly, J. L., and Farley, J. W. (1998) Capillary electrophoresis/laser-induced fluorescence detection of fluorescein as a groundwater migration tracer. Electrophoresis 19, 2252-2256.
15. Brumley, W. C., and Gerlach, C. L. (1998) Capillary electrophoresis/laser-induced fluorescence in groundwater migration determination. Amer. Laboratory 31, 45-49.
16. W. C. Brumley and J. W. Farley, Determining eosin as a groundwater migration tracer by capillary electrophoresis/laser-induced fluorescence using a multiwavelength laser@, Electrophoresis, 24, 2335-2339 (2003).
17. Brumley, W. C., Grange, A. H., Kelliher, V.,  Patterson, D. B., Montcalm, A., Glassman, J., and Farley, J.  (2000) Environmental screening of acidic compounds based on CZE/LIF detection with GC/MS and GC/HRMS identifications. JAOAC International 83, 1059-1067.
18. Brown, R.S., Luong, J.L.T., Szolar, O., Halasz, H.J., and Hawari, J.A. (1996) Cyclodextrin-modified capillary electrophoresis: Determination of polycyclic aromatic hydrocarbons in contaminated soils. Anal. Chem. 68, 287-292.
19. Yan, C., Dadoo, R., Zhao, H., Zare, R.N., and Rakestraw, D.J. (1995) Gradient elution in capillary electrochromatography. Anal. Chem., 67 2026-2029.
20. Dabek-Zlotorzynska, E., Piechowski, M., Liu, F., Kennedy, S., and. Dlouhy, J. F (1997) Routine determination of major ions in atmospheric aerosols by capillary electrophoresis. J. Chromatogr. A 770, 349-359.
21. Dabek-Zlotorzynska, E. and Maruszak, W.  (1998) Determination of dimethylamine and other low-molecular-mass amines using capillary electrophoresis with laser-induced fluorescence detection. J. Chromatogr. B 714, 77-85.
22. Nordén, M. and Dabek-Zlotorzynska, E. (1996) Study of metal-fulvic acid interactions by capillary electrophoresis J. Chromatogr. A 739, 421-429.
23. Flaherty, S., Wark, S., Street, G., Farley, J.W., and Brumley, W. C. (2002) Investigation of CE/LIF as a tool in the characterization of sewage effluent for fluorescent acidics:  Determination of salicylic acid", Electrophoresis 23 (14), 2327-2332.
24. Jung, M. and Brumley, W. C. (1995) Trace analysis of fluorescein-derivatized phenoxy acid herbicides by micellar electrokinetic chromatography with laser-induced fluorescence detection. Chromatogr. A 717, 299-308.
25. El Rassi, Z., Mechref, Y., Postlewait, J., and Ostrander, G. K. (1997) Capillary Electrophoresis of Carboxylated carbohydrates. III. Selective Precolumn Derivatization of Glycosaminoglycan Disaccharides with 7-Aminonaphthalene-1, 3-Disulfonic Acid Fluorescing Tag. Anal. Biochem. 244, 283-290.
26. K. R. Rogers, A. B. Apostol, and W. C. Brumley, (2000) Capillary electrophoresis (CE) immunoassay format for phenoxyacid herbicides. Anal. Lett. 33, 443-453.
27. Wall, W., El Rassi, Z. (2001) Electrically driven microseparation methods for pesticides and metabolites: V. Micellar electrokinetic capillary chromatography of aniline pesticidic metabolites derivatized with fluorescein isothiocyanate and their detection in real water at low levels by laser-induced fluorescence. Electrophoresis 22, 2312-2319.
28. Brumley, W. C., and Kelliher, V. (1997) Determination of aliphatic amines in water using derivatization with fluorescein isothiocyanate and capillary electrophoresis/laser-induced fluorescence detection. J. Liq. Chromatogr. 20, 2193-2205.
29. Kok, S.J., Hoornweg, G., de Ridder, T., Brinkman, U., Velthorst, N.H., Gooijer, C., (1998) Generation of 275.4-nm UV output from a large-frame argon-ion laser for fluorescence detection in capillary electrophoresis.  J. Chromatogr. A 806, 355‑360.
30. Kok S. J., Kristenson, E. M., Gooijer, C., Velthorst, N. H., and Brinkman, U. A. Th.  (1977) Wavelength-resolved laser-induced fluorescence detection in capillary electrophoresis: naphthalenesulphonates in river water. J. Chromatogr. A 771, 331-341.
31. Kok, S.J.,  Isberg, I. C. K., Gooijer,  C.,  Brinkman,  U. A. Th., and  Velthorst,  N. H. (1998) Ultraviolet laser-induced fluorescence detection strategies in capillary electrophoresis: determination of naphthalene sulphonates in river water. Analytica Chimica Acta 360, 109-118. [32] Holcomb, M. and Thompson, H.C. (1996) Analysis of fumonisin B1 in rodent feed by capillary electrophoresis with fluorescence detection of the FMOC derivative. J. Capillary Electrophoresis, 3(4),
32. Cole, R.O., Holland, R.D., and Sepaniak, M.J. (1992) Factors Influencing Performance in the Rapid Separation of Aflatoxins by Micellar Electrokinetic Capillary Chromatography," Talanta, 39(9), 1139-1147.
33. de Geus, H. - J., de Boer, J. and Brinkman, UA.Th. (1996). Multidimensionality in gas chromatography. Trends Anal. Chem. 15, 398-408.
34. Bushey, M.M. and Jorgenson, J.W. (1990) Automated instrumentation for comprehensive two-dimensional high performance liquid chromatography of proteins.  Anal. Chem. 62, 161-167.

Figure legends

Fig. 1.   Schematic diagram of an optical bench with CE/LIF experiment and data system acquisition.  Laser light is aligned using mirrors, possibly attenuated or subjected to a notch filter at the excitation wavelength (e.g., 325 nm), focused using a microscope objective onto the capillary at the window.  Fluorescent light is collected at 90º  to the beam, collected and refocused by a second objective to a point and passes through a slit and band pass filters to the photomultiplier tube for amplification and conversion to a digital signal for the computer.

Fig. 2.  CE/LIF detection of fluorescein (MT=3.58 min) (excitation 488 nm, emission band pass centered at 520 nm) as a groundwater tracer in an actual study.  Groundwater was exposed to the charcoal pad, eluted with solvent, and quantitated at 62 ppt in the extract.  Internal standard erythrosin B appears at MT=3.30 min and the EO flow disturbance near 2.0 min. Separation used a 57 cm capillary (0.075 mm ID, 50 cm to the detector) with 40 mM borate buffer and 30 kV voltage in a Beckman P/ACE 5000 instrument.