Trace Organic Analysis
Determining eosin as a groundwater migration tracer by capillary electrophoresis/laser-induced fluorescence using a multiwavelength laser
Published in Electrophoresis, 24, 2335-2339 (2003)
[note: minor content and formatting differences exist between this web version and the published
William C. Brumley1 and John W. Farley2
1U.S. Environmental Protection Agency, Environmental Sciences Division, P.O. Box 93478, Las Vegas, NV 89193-3478
2Department of Physics, University of Nevada-Las Vegas, P.O. Box 454002, Las Vegas, NV 89154-4002
Correspondence: William C. Brumley, U. S. EPA, P. O. Box 93478, Las Vegas, NV 89193-3478; tel: 702-798-2684; fax: 702-798-2142; E-mail: firstname.lastname@example.org
|HPLC||high performance liquid chromatography|
|QA/AC||quality assurance/quality control|
|RCRA||Resource Conservation and Recovery Act|
|Superfund||created by the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA)|
|ppt||parts per trillion|
Measurements for determining of the path of groundwater migration remain an important tool in the overall assessment of environmental processes and transport of pollutants. This paper examines a multiwavelength laser for the determination of eosin, a groundwater tracer, using capillary electrophoresis/laser-induced fluorescence (CE/LIF) at excitation wavelength 514.5 nm. Eosin was one of four dyes used in a study of adjacent RCRA and Superfund sites that routinely relied on spectrofluorimetry for determination as we have previously reported. However, the improved specificity of CE/LIF is further illustrated in this work applied to the analysis of adsorbent pads placed in monitoring wells after dye injection and flushing from injection wells. The multiwavelength laser provided the capability to analyze for several dyes with one laser. The advantages/disadvantages of CE/LIF versus spectrofluorimetry are discussed. Spectrofluorimetry is fast and sensitive and will likely continue to be the primary workhorse technique. CE/LIF could provide confirmation when greater specificity is needed in a regulatory context.
Groundwater migration remains an important contributor in determining the distribution and fate of environmental pollutants originating from various waste sites or in understanding fate and transport [1-3]. Groundwater tracers are often used to determine the flow of groundwater. The tracers can be fluorescent dyes, in which case the use of synchronous scanning spectrofluorimetry is the technique of choice to determine the tracers in water or other samples. The U.S. EPA National Exposure Research Laboratory (Las Vegas) is interested in developing and applying analytical tools that can strengthen the regulatory application of analysis. Among those tools under investigation is capillary electrophoresis/laser-induced fluorescence detection (CE/LIF), which is particularly well-suited to the analysis of water-soluble and fluorogenic compounds.
Fluorescent dyes are a convenient choice for tracers because of the ease of sensitive detection. Spectrofluorimetry, high performance liquid chromatography (HPLC)/UV or fluorescence detection, and capillary LC/fluorescence detection have also been used [2-6]. For HPLC or capillary LC, the retention of dye analytes is enhanced using ion-pairing techniques. Logically, applications of capillary electrophoresis/laser-induced fluorescence (CE/LIF) detection should be ideal for the determination of anionic (or cationic) dyes. In free-zone CE, the dyes are separated very simply on the basis of their mobilities in aqueous buffers. Among several reports on CE/LIF detection techniques are two early papers based on high-sensitivity detection of fluorescent dyes [7,8] and more recent work applied directly to groundwater migration studies [9-12]. General reviews of environmental applications of CE and sample handling were published , and more recently updated reviews appear regularly [14-15] Additional environmental applications of CE/LIF have also appeared [16-17].In this work, we present the use of a multiwavelength laser and application of CE/LIF detection for determining eosin as a tracer compound for groundwater migration. A comparison is made between the performance of CE/LIF and that of traditional spectrofluorimetry. Attention is drawn to the role of the separation and the use of a multiwavelength laser in determining dyes adsorbed to receptor pads placed in monitoring wells. To our knowledge, this is the first application of CE/LIF to groundwater migration using eosin dye and a multiwavelength laser. Previously, we had reported studies of tinopal  and fluorescein  as groundwater tracers where single wavelength lasers were used to generate the excitation wavelength.
2 Materials and Methods
Compounds were obtained as follows: (a) sodium hydroxide 98.9%, (Fisher Scientific, Pittsburgh, PA); (b) hydrochloric acid ACS grade, 36.5-38% (Fisher Scientific); (c) methanol ACS/HPLC grade (Fisher Scientific); (d) acetone AR grade (Fisher Scientific); (e) sodium tetraborate decahydrate ACS grade (Aldrich); (d) dyes eosin Y, erythrosin B, fluorescein, and eosin B were 90% or greater (Aldrich Chemical Co.). Solutions were made up as follows:
(a) 6N NaOH 100 mL batches were freshly prepared from 24.3 g of the solid NaOH made up to 100 mL with deionized water. 0.1N NaOH solution was prepared from the 6N solution by serial dilution.
(b) Borate Buffer 20 mM borate buffer was prepared from 0.1907 g of sodium tetraborate decahydrate and dissolved in deionized water and diluted to 100 mL.
Other materials were prepared as follows: Receptor Pads prepared by the on-site contractor and consisted of fiberglass mesh filled with coconut charcoal and weighted to remain near the bottom of the monitoring well.
2.2 Capillary Electrophoresis
A laboratory-built CE/LIF instrument previously described  was used for all electrophoretic determinations reported here. The instrument was fitted with a capillary 58 cm in total length (50 cm from the origin to the detector window) and 75 Ám inside diameter. Fluorescence determination was accomplished with an Ar/Kr ion laser operated at 514.5 nm line of the laser and detection using a photomultiplier tube at 90 deg to the excitation beam and preceded by two band-pass filters (Omega Optical, Battlesboro, VT, 540DF20) and an entrance slit. Unless otherwise noted, electrophoresis was carried out at 18 kV. The temperature of the capillary was ambient (ca. 25C). The capillary was equilibrated with running buffer for two minutes prior to beginning of an experiment, and washed for two minutes with 0.1N NaOH solution and DI 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 supplemental software were further normalized by dividing them by the area of the peak of the fluorescein internal standard. This computation was necessary to correct for the small variations in injection volumes resulting from the pressure injections typically of 7 sec duration (nominally 30 nL). The eosin standard exhibited three peaks as well as an early-eluting impurity.
2.2.1 Calibration Curve
A regression analysis was carried out on the ratio of corrected areas (eosin corrected area divided by internal standard corrected area) versus the ratio of eosin concentration to internal standard concentration (1 X 10-6 M fluorescein) resulting in a 6-point calibration curve. Both unweighted and weighted regressions were considered and resulted in an r2 of 0.995 and 0.997 respectively. The equation of the line was used to calculate the concentrations of eosin in the samples and check standards run during the course of analysis. The concentration of eosin was calculable based on the known volume and concentration of internal standard added to the sample and the known volume/mass of the sample being analyzed.
2.2.2 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 migration time (MT) variations due to changes in electroosmotic (EO) flow. Deviations from the calibration curve greater than ▒15% caused rerunning of standards, construction of a new calibration curve, or replacement of the capillary.
The MT 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 analyte and fluorescein. Using MT correction based on the internal standard , the expected and measured MT of the eosin peak fell within 0.3%. The position of both the internal standard and eosin could also be estimated from the position of the EO flow disturbance. Since the internal standard, fluorescein, eluted just after the three eosin peaks, the position of the analyte was deduced with great certainty. 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 CE/LIF.
2.3 Groundwater Injection Studies
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 by an on-site contractor. 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 separate samples of water and receptor pads. Here, we report results only for the eosin tracer in all 22 samples consisting of receptor pad extracts.
2.4 Receptor Pad Sample Handling
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 (v:v:v-propanol:water:concentrated ammonium hydroxide) hereafter designated ES. In this study, however, only 0.5 g of charcoal from the receptor was extracted in 5 mL of ES to minimize solvent usage and because the samples could be concentrated when using CE/LIF analysis. The extract was evaporated to dryness and redissolved in 50:50 acetone:water and internal standard added. Results for receptor analysis are reported as ppt-levels in the 5-mL extract of the pads for 0.5 g samples.
Fluorescence emission spectra were recorded using a Spex Fluorolog 2, which is a double monochrometer instrument (excitation and emission) with 450 W Xe lamp (Instruments S.A., Edison, NJ). Groundwater studies used maximum sensitivity (1-cm path-length cells) with 5.0-mm slits (8.5-nm bandpass) and a synchronous scan (1 sec/nm, 1-sec integration time) with a Stokes shift of 20 nm (optimized for the visible fluorescent dye tracers).
3 Results and Discussion
3.1 Migration Studies
Fluorescent dyes were injected into wells at the RCRA site, and monitoring wells located at the nearby Superfund site were sampled regularly by means receptors placed in the monitoring wells for periods of one to several days. Eosin was found to have an absorption maximum near the 514.5 nm line of the Ar/Kr mixed gas laser. The fluorescence emission maximum of eosin for excitation at 514.5 nm was found to be near 535 nm so that LIF and spectrofluorimetry were operating in a roughly equivalent manner.
The fluorescein internal standard eluted right after the third eosin peak at 5.23 min. An impurity (multiplet) also is seen with a MT of 3.81 min that arises from the eosin standard. We are not aware of reports on multiple peaks for the eosin standard. In an attempt to see if the peaks resulted from an erythrosin (structurally identical to eosin but containing four iodo groups instead of 4 bromo groups) impurity, we found that erythrosin is indistinguishable from the first peak of eosin by CE. Thus, the peaks may arise from positional isomers of eosin (carboxylate group) but we have not established this experimentally. We note that all three dyes: fluorescein, eosin, and erythrosin are based on the xanthene structure. The power of coupling a separation with fluorescence detection is an obvious advantage of CE/LIF, which in this instance has apparently newly revealed the multicomponent nature of the dye.The choice of fluorescein as internal standard deserves some discussion. Normally, the choice of internal standard does not include one of the analytes. One of the difficulties is finding an appropriate internal standard that fluoresces but does not coelute with the analyte. However, in this case the parameters of the experiment have been altered from those where fluorescein was analyzed. Both the excitation and emission regions have changed. Validation has been accomplished by establishing that fluorescein was not in the samples prior to the determination of eosin or was at most three orders of magnitude below the spiking level of internal standard. Thus, the quantitative results are unaffected. Using the multiwavelength laser, it is a straightforward matter to check the fluorescein internal standard response and migration time using the 488 nm line of the laser (and a new set of emission filters) without having to change lasers or optical bench alignment. A calibration curve is shown in Figure 2.
Figure 2. Calibration line for eosin versus internal standard (fluorescein) (r2=0.995).
In practice, little use of the quantitative number is made in groundwater
migration studies since the detection of the dye is considered the primary
finding relative to migration direction. Thus, the certainty of the finding
of the dye provided by CE/LIF, where the MT enhances the qualitative identification,
strengthens the conclusion. This is especially useful in a regulatory
setting where a preponderance of evidence is sought.
3.2 Detection Levels Achieved in Receptors
In agreement with the previous results from synchronous scans from spectrofluorimetry, no eosin was detected in any of the 22 pads. However, as illustrated in Figure 3, one or more additional peaks were detected by CE/LIF (in this case at 5.32 min) that potentially could be capable of giving a false response via spectrofluorimetry.
Thus, the addition of the separation has demonstrably improved the specificity of the determination, eliminating false positives. As an illustration of an achievable lower detection level, a spiked sample at 100 ppt was detected and quantified by CE/LIF as 47 ppt (Fig. 4); differences in migration time versus Fig. 1 are due to use of a different capillary. CE/LIF (ex 514.5 nm) of receptor pad extract with fluorescein (5.04 min) internal standard and coextractive peak at 5.32 min. Separation conditions: 0.075 mm ID X 58 cm capillary (window 8 cm from end); running buffer was 20 mM borate buffer (pH 9.2); separation voltage 18 kV
For reasons not entirely clear to us, the detection limit of eosin was not as low as expected (ca 1 X 10-9 M rather than 1 X 10-10 M and was inferior to the approximately 10-10 M detection limit of spectrofluorimetry in DI water and the 3 ppt level in actual samples ).
3.3 Variations in Migration Times
Under free-zone electrophoresis, variations in MTs of target ions are principally due to variations in the EO flow. Corrections to this variation were carried out  and corrected migration times for eosin generally agreed within 0.3 % for predicted versus observed MTs for work carried out on the same capillary. Changing the capillary often leads to a new set of migration times. Variations outside this range on a given capillary may be due to a number of factors (e.g., adsorption or partially blocked capillary) or could indicate that the peak assignment is questionable. The presence of the internal standard response serves as the primary QC aspect of the analysis together with the behavior of standards.
3.4 Advantages/Disadvantages of CE/LIF Versus
CE/LIF offers greater specificity in determining dyes and eliminating background signal levels. The ease of sample handling favors spectrofluorimetry because no filtration is required and all four dyes could be determined in a single run. With CE/LIF, a change in excitation wavelength is required (but not a change in laser or alignment) as well as a change in emission filters. With the single multiwavelength laser, a variety of dyes could be detected because of the large number of lines available. However, spectrofluorimetry will likely remain the workhorse technique since it is more widely available. Availability of a CE instrument and proper laser excitation sources may be current drawbacks for some laboratories. Consideration of using CE/LIF as a confirmatory technique for spectrofluorimetry is recommended, especially when results are not clear cut or the regulatory context requires confirmation. Spectrofluorimetry is capable of high throughput with about 5 min per sample required compared with about 10 min per sample for CE but only one wavelength. An advantage of commercial CE instrumentation is the ability to use an autosampler to run samples unattended. Spectrofluorimetry is a straightforward technique that can be used with minimal training and is recommended for screening purposes. Cost of instrumentation, however, actually favors the CE/LIF system. Other dyes are accessible via a variety of laser sources, some of which are HeNe lasers that are relatively inexpensive.
4 Concluding Remarks
CE/LIF detection offers a sensitive, specific approach to groundwater migration studies involving fluorescent tracer dyes. The analytical problem of detecting fluorescent tracer compounds (ionic analytes) in water samples is perfectly matched to separations involving free-zone CE. The detection limits afforded by CE/LIF are as good or better than those of traditional spectrofluorimetry, while providing increased specificity because of the separation based on ion mobilities. Future studies of groundwater migration should be planned with this capability in mind. Choice of tracer dyes can be based on the availability of lasers matched to the absorption maximum of appropriate dyes.
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development, funded and performed the research described here. This work has been subjected to the Agency's peer review and has been approved as an EPA publication. The U.S. Government has the right to retain a nonexclusive, royalty-free license in and to any copyright covering this article. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
 Smart, P. L., Laidlaw, I.M.S., Water Resources Res. 1977, 13, 15-33.
 Davis, S. N., Campbel, D. J., Bentley, H. W., Flynn, T. J., Ground Water Tracers, 1985, Natl. Water Well Assoc., Worthington, OH; Cooperative agreement CR-810036, U.S. Environmental Protection Agency, Robert S. Kerr Environmental Research Laboratory, Ada, OK 74820.
 Aulenbach, D. B., Bull, J. H., Middlesworth, B.C., Ground Water, 1978, 16,149-157.
 Laane, R.W.P.M., Manuels, M. W., Staal, W. Water Res., 1984, 18, 163-165.
 Van Soest, R.E.J., Chervet, J. P., Ursem, M., Suijlen, J. M., LCGC International, 1996, September, 586-593.
 Jones, T. L., Master's Thesis: the Detection and Evaluation of Fluorescent Brighteners as Organic Tracers in Alkaline Groundwaters, Department of Chemistry, University of Nevada-Las Vegas, Las Vegas, NV, December, 1991.
 Timperman, A. T., Khatib, K., Sweedler, J. V., Anal. Chem., 1995, 67, 139-144.
 Chen, D. Y., Dovichi, N. J., Anal. Chem., 1996, 68, 690-696
 Ferguson, P., Undergraduate Thesis: The Use of Spectrofluorimetry and Capillary Electrophoresis/Laser-Induced Fluorescence for the Determination of Fluorescent Dyes in Groundwater Migration Studies, Department of Environmental Studies, University of Nevada-Las Vegas, NV, December, 1997.
 Brumley, W. C., Ferguson, P. L., Grange, A. H., Donnelly, J. L., Farley, J. W., J. Cap. Electrophor., 1996, 3, 295-299.
 Ferguson, P. L., Grange, A. H., Brumley, W. C., Donnelly, J. L., Farley, J. W., Electrophoresis, 1998, 19, 2252-2256.
 Brumley, W. C., Gerlach, C., Am. Lab., 1999, 31, 45-49.
 Brumley, W. C., LCGC, 1995, 13, 556-568.
 Dabek-Zlotorzynska E; Aranda-Rodriguez R; Keppel-Jones K., Electrophoresis, 2001, 22, 4262-4280.
 Tegeler T; El Rassi Z., Electrophoresis, 2001, 22, 4281-4293.
 Brumley, W. C., Grange, A. H., Kelliher, V., Patterson, D. B., Montcalm, A.,
Glassman, J., Farley, J. W., JAOAC International, 2000, 83, 1059-1067.
 Flaherty, S., Wark, S., Street, G., Farley, J. W., Brumley, W. C., Electrophoresis, 2002, 23, 2327-2332.
 Brumley, W. C., Brownrigg, C. M., J. Chromatogr., 1993, 646, 377-389.