Anal. Chem. 1997, 69, 1726-1731
Separation and Determination of Polycyclic Aromatic Hydrocarbons by Solid Phase Microextraction/Cyclodextrin-Modified Capillary Electrophoresis An-Lac Nguyen and John H. T. Luong*
Biotechnology Research Institute, National Research Council Canada, Montreal, Quebec, Canada, H4P 2R2
A sensitive method for the determination of polycyclic aromatic hydrocarbons (PAHs) by solid phase microextraction coupled with cyclodextrin (CD)-modified capillary electrophoresis (CE) using UV detection has been developed. A glass fiber was prepared and used for absorbing 16 EPA priority PAHs from diluted samples until equilibrium was reached. After the glass fiber was connected to a separation capillary via an adapter, the absorbed analytes were directly released into the CE buffer stream, and electrophoretic separation was effected using a 50 mM borate, pH 9.2, buffer containing 35 mM sulfobutyloxy-β-CD, 10 mM methyl-β-CD, and 4 mM r-CD. Separation was effected since neutral PAHs differentially partitioned between the neutral and charged CD phases. Under 30 kV applied potential, separation was achieved in less than 15 min with high resolution and number of theoretical plates. Pyrene as low as 8 ppb was detected, while the highest limit of detection was 75 ppb for acenaphthene. Very satisfactory reproducibility with respect to migration time and peak area was obtained for repetitions using the same separation capillary and adapter, where only the extraction fiber was discarded after each analysis. Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous environmental pollutants which present a potential health concern because of the toxicity, mutagenicity, and carcinogenicity of these substances in animals. Therefore, development of inexpensive, simple, and sensitive analytical methods is of utmost importance to detect these water-insoluble toxicants. Capillary electrophoresis (CE) has emerged as one of the most efficient methods for separation of charged components in mixtures.1 Modified CE methods which can separate uncharged and hydrophobic molecules such as PAHs have also emerged, most notably micellar electrokinetic capillary chromatography2 and cyclodextrin (CD)modified CE using a mixture of neutral and charged CDs.3,4 Separation will be effected, provided neutral molecules differentially partition between the buffer and micelle or CD phases. As an example, mixtures of negatively charged sulfobutyloxy-β-CD (1) Sepaniak, M. J.; Powell, A. C.; Swaile, D. F.; Cole, R. O. In Capillary Electrophoresis, Theory and Practice; Grossman, P. D., Colburn, J. C., Eds.; Academic Press: San Diego, CA, 1992; pp 159-189. (2) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1984, 56, 111. (3) Szolar, O. H. J.; Brown, R. S.; Luong, J. H. T. Anal. Chem. 1995, 67, 3004. (4) Brown, R. S.; Luong, J. H. T.; Szolar, O. H. J.; Halasz, A.; Hawari, J. Anal. Chem. 1996, 68, 287.
1726 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
(SBβCD) and neutral methyl-β-CD (MβCD) were able to separate 16 Environmental Protection Agency (EPA) priority PAHs4 as well as other aromatic hydrocarbons with high efficiency, resulting in remarkable resolution of very similar isomers such as benzo[a]pyrene and benzo[e]pyrene.3 Satisfactory separation of all 16 PAHs was achieved in under 20 min, with efficiencies for all components greater than 105 theoretical plates. The separation was sufficient to analyze several PAHs and a variety of other compounds in soil extracts.4 Despite the great resolution power, the popularity of CE is hampered by its sensitivity, which is inherent in the small path length (the same as the capillary internal diameter, i.d.) when on-line absorbance detectors are used. Attempts have been made to alleviate this limitation by using commercially available bubble cell capillaries3,4 and Z-shape high sensitivity optical cells, but only 4- to 5-fold improvement in sensitivity can be realized. For fluorescent compounds, sensitivity could also be significantly improved with UV/visible laser-induced fluorescence detection,5 but laser stability (e.g., HeCd laser) and the exorbitant cost of UV lasers still represent a major hurdle. If non-fluorescent compounds are of interest, they must undergo tedious and complicated derivatization procedures that are difficult to perform in small volumes or at low concentrations.5 In view of this, detection sensitivity has become the “bottleneck” in CE analysis of environmental pollutants such as benzene and its derivatives, polychlorinated phenols, and chlorinated dioxins and furans. Solid phase microextraction (SPME) has been recently developed and successfully applied in conjunction with gas chromatography (GC) to analyze very diluted samples.6 The basic concept resides in absorbing the target analyte from the sample matrix onto a thin coat of polymer that is deposited on a small glass fiber. The analyte is absorbed until equilibrium is reached in the system, and up to 100 ng of an analyte can be absorbed onto the coated fiber.6 The absorbed analytes are then directly released into the GC carrier gas stream. SPME can be regarded as a combined sampling and sample concentration technique, requiring very simple and low-cost devices. This technique has been developed for several applications, including the determinations of benzene derivatives,7,8 polyaromatic hydrocarbons and (5) Nie, S.; Dadoo, R.; Zare, R. N. Anal. Chem. 1993, 65, 3571. (6) Hawthorne, S. B.; Miller, D. J.; Pawliszyn, J.; Arthur, C. L. J. Chromatogr. A 1992, 603, 185. (7) Arthur, C. L.; Killam, L. M.; Motlagh, S.; Lim, M.; Potter, D. W.; Pawliszyn, J. Environ. Sci. Technol. 1992, 26, 979. (8) Potter, D. W.; Pawliszyn, J. J. Chromatogr. A 1992, 625, 247. S0003-2700(96)00986-9 CCC: $14.00
© 1997 American Chemical Society
polychlorinated biphenyls (PCBs),9 chlorinated hydrocarbons,10 and tetraethyllead as well as inorganic lead in water.11 This article presents the first application of SPME together with cyclodextrin-modified CE equipped with UV detection for separation and sensitive analysis of PAHs. The method used a negatively charged cyclodextrin, SBβCD, and two neutral cyclodextrins, MβCD and R-CD, for separation of 16 EPA priority PAHs. The sensitivity and reproducibility of the SPME-CE procedure will be also investigated and discussed. EXPERIMENTAL SECTION Materials. Hydroxypropyl-β-CD (HPβCD), with an average degree of substitution of seven hydroxypropyls per cyclodextrin (American Maize, Hammond, IN), and SBβCD, with an average degree of substitution of 4 (Applied Biosystems Division, PerkinElmer, Foster City, CA; kindly given by Dr. John Stobaugh, University of Kansas), were used as received. MβCD, with a degree of substitution of 12.6 methyl groups per cyclodextrin, R-CD, and γ-CD were obtained from Aldrich (Milwaukee, WI). All other chemicals were products of Sigma (St. Louis, MO). Water was obtained from a Zenopure Quadra 90 (Zenon Environmental, Burlington, ON, Canada) cartridge filtration system, with a specific resistivity greater than 15 MΩ‚cm. Sample Preparation. A solution of 16 priority PAHs (EPA 610) was purchased from Supelco (Bellefonte, PA). This standard solution in methanol/methylene chloride (1:1) was diluted 40fold in methanol (designated as sample 40×) for analyses without the SPME procedure. A 40× sample was further diluted 100fold in water (designated as sample 4000×) for analyses with the SPME procedure. In sample 4000×, the lowest and highest analyte concentrations were 25 and 500 ppb (for pyrene and acenaphthylene, respectively). Samples for determination of detection limits were prepared from 1 mg/mL stock solutions of individual PAHs in methanol. The stock solutions were diluted in water to obtain solutions of desired concentrations. CE Equipment and Data Treatment. All capillary electropherograms were obtained from a P/ACE 5000 (Beckman, Fullerton, CA) instrument. Separations were done at 30 kV using a 50 µm i.d., 350 µm o.d. fused silica capillary with an inlet-todetector length of 50 cm and a total length of 57 cm, unless otherwise indicated. In all cases, the absorbance was monitored at 254 nm with a Beckman UV detector. Peak identification was performed by spiking with individual components. Electropherograms were also analyzed by determining the number of theoretical plates (N) for each peak in the electropherograms, according to the equation N ) 5.54(t/w1/2)2, where t and w1/2 represent the migration time of the analyte and the full peak width at half-maximum, respectively.12 The Beckman Gold software furnished with P/ACE calculated N from each electropherogram on command. The resolution (R) of one peak from the preceeding peak was calculated as 1.18(t2 + t1)/(w1/2,1 + w1/2,2). Solid Phase Microextraction. An optical fiber (Polymicro Technologies Inc., Phoenix, AZ) was cut into 3 cm pieces, and then about 1.5 cm of each piece was dipped in dimethylformamide (9) Potter, D. W.; Pawliszyn, J. Environ. Sci. Technol. 1994, 28, 298. (10) Chai, M.; Arthur, C. L.; Pawliszyn, J.; Belardi, R. P.; Pratt, R. P. Analyst 1993, 118, 1501. (11) Gorecki, T.; Pawliszyn, J. Anal. Chem. 1996, 68, 3008. (12) Anigbogu, V. C.; Copper, C. L.; Sepaniak, M. J. J. Chromatogr. A 1995, 705, 343.
Figure 1. Attachment for solid phase microextraction.
for 3 min so that the swollen nylon cladding of this portion could be stripped with a pair of tweezers. After stripping of the nylon cladding, the fiber still retained a coat of poly(dimethylsiloxane) (PDMS) over the glass core. The stripped pieces were soaked in deionized water overnight, and then the stripped portions were cut to 1.2 cm. Absorption of analytes was effected by positioning the partially stripped fiber piece at the side of a centrifuge tube (0.5 mL) containing 0.1 mL of the sample. Care was exercised to allow one end of the piece to reach the center of the tube bottom, and the absorption time was at least 2 h to reach equilibrium between the fiber and the aqueous solution. It should be noted that PAHs such as naphthalene, anthracene, benz[a]anthracene, and benzo[a]pyrene were shown to reach equilibrium with this type of fiber within 1-2 h.9 Coincidentally, 0.1 mL filled the centrifuge tube to a depth of 1 cm; therefore, analytes were absorbed in the coating of 1 cm of the coated fiber. Since the stripped portion of each piece was cut to 1.2 cm, the nylon cladding of the intact portion (not stripped) did not contact the sample solution. A fiber piece with absorbed analytes was soaked in water for at least 30 min, to wash off any surface-attached material, before being installed on the separation capillary. CE with SPME Attachment. The P/ACE instrument uses cassettes to mount the separation capillaries, leaving about 4 cm of capillary at each end. In creating the attachment for installing the fiber, 1.5 cm was cut from the inlet end of the separation capillary, and then a piece of heat-shrinkable tubing (2 cm long, designed to fit 0.38 mm capillary od, MicroQuartz Science, Phoenix, AZ) was slipped over the capillary end to cover about 1 cm. A short capillary segment (1.5 cm long, 180 µm i.d., 380 µm od, MicroQuartz) was inserted into the free portion of the tubing, and then a gentle stream of warm air was directed at the tubing to form an air-tight coupling between the larger internal diameter segment and the separation capillary. One end of the larger internal diameter segment was well polished by the supplier, and if the separation capillary was properly cut and polished, the coupling would approach zero dead volume (Figure 1). After an SPME attachment is coupled to a cassette-mounted capillary, the integrity of the heat-shrunk junction must be tested. Testing was conducted by using the cassette to obtain an electropherogram with a standard sample solution and comparing that with an electropherogram previously obtained for the same Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1727
sample, using a capillary (with the same total length) without the SPME attachment. Any improperly sealed junction would cause current failure during the electrophoresis run, while significantly different electropherograms (compared with that obtained without the attachment) would indicate unacceptable dead volume at the junction. After passing the test for junction integrity and dead volume, the SPME attachment-bearing capillary was removed from the P/ACE system, ready for insertion of the analyte-bearing fiber (see preparation above). Such an analyte-bearing fiber had a 1.2 cm uncladded portion bearing the analytes and a 1.5 cm nylon cladded portion. The nylon cladded portion (necessary to allow easy handling during the analyte absorption) was cut to leave only 0.5 cm before the fiber was inserted into the attachment. Since the fiber has a 150 µm glass core, it could be easily inserted into the 180 µm i.d. capillary segment, and a microscope greatly facilitated the procedure. After fiber insertion, the cassette was installed in the P/ACE system to fill the capillary, followed by 3 s of methanol injection. The system was left idle for 5 min to allow redissolution of the analytes. Potential was then applied to perform electrophoresis in the normal manner. After each completed run buffer was forced into the capillary in the reverse direction to eject the fiber, and the cassette was then removed from the instrument, ready for installation of another fiber. Time saving and better use of machine time were achieved by using two cassettes, each having an attachment. Fiber insertion was thus performed with one cassette while the other was on the machine. Careful planning with alternating use of two cassettes would allow replications with the same cassette for better reproducibility (see Results). It is worth noting that the SPME attachment was installed at the inlet end of the capillary, while the P/ACE system is always operated with photometric detection (UV or fluorescent) 7 cm from the outlet. The SPME attachment, with or without the glass fiber, had no bearing on the detector. Pieces of optical fibers were convenient starting materiel to obtain glass fiber segments, uniformly coated with poly(dimethylsiloxane) (PMDS, for analyte absorption), each of which also had a nylon-coated portion for easy handling. The optical properties of the fiber were not a required feature. The PMDS-coated glass fiber (without nylon cladding) was also tested in view of eliminating the dimethylformamide stripping step, but the fiber segments without the nylon cladded portion were too fragile. Safety. Several PAHs used in this study are suspected carcinogens, and caution must be exercised when working with these chemicals. Stock solutions of PAHs were handled in a ventilated hood and stored in closed containers. Disposable latex gloves were worn while working with PAHs, and care was taken to dispose of waste solutions properly. RESULTS AND DISCUSSION Improvement of Cyclodextrin-Modified CE. Previously reported results4 showed that the 16 EPA priority PAHs could be satisfactorily separated using a running buffer containing 35 mM SBβCD and 15 mM MβCD in 50 mM borate, pH 9.2. However, under this operating condition, eight PAHs exhibited very close migration times to form a cluster, whereas a second cluster was formed by another three compounds (Figure 2a). In an attempt to improve separation, SBβCD was increased to 70 mM while decreasing MβCD to 10 mM to delay migration of most PAH components. With this buffer, 13 PAHs were well resolved, but 1728 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
Figure 2. Electropherograms obtained with 50 mM borate buffer pH 9.2. (a) 35 mM SBβCD, 15 mM MβCD, sample 20×. (b) 70 mM SBβCD, sample 40×. (c) 35 mM SBβCD, 10 mM MβCD, 4 mM R-CD, sample 40×. Peaks in (c) correspond to the order presented Table 1. Each number in (a) corresponds to the component identified in Table 1.
unfortunately four others migrated as one new cluster (Figure 2b). It should also be noted that the migration order was significantly altered. As discussed previously,3 each analyte should display a particular partitioning coefficient toward SBβCD and MβCD, and separation was primarily governed by differences in partition coefficients. The acquired experimental results clearly illustrated that any change in the concentration ratio will have pronounced effects on the migration of the analytes, and this behavior can be exploited to resolve the components of interest. The last peak, i.e., no. 16 or benzo[ghi]perylene, appeared to interact strongly with SBβCD and/or the capillary wall, and in some cases it did not appear in the electropherogram, even after 20 min of separation. Evidence of microprecipitation, i.e., spikes, was also observed in certain runs. Attempts to introduce γ-CD (up to 10 mM) and hydroxypropylβ-CD (up to 10 mM) as well as urea (up to 5 mM) in the running buffer containing 35 mM SBβCD and 15 mM MβCD did not result in any improvement in separation resolution. The negative results obtained with γ-CD were somewhat surprising since one may expect that this eight-glucose cyclodextrin, with its fairly large cavity (7.9 Å),13 should be able to interact with most of the 16 PAHs to improve separation. In contrast, Figure 2c showed that better resolution was effected with 4 mM R-CD, which, in general, was not expected to interact with the PAHs in view of its smaller (13) Szejtli, J. Cyclodextrin Technology; Kluwer Academic Publishers: Dordrecht, 1988.
Table 1. Comparison of the Separation Efficiency of the Buffer Medium in the Presence of 4 mM r-Cyclodextrin plate no. × 103 (N) PAH
without R-CD
with R-CD
1. dibenz[a,h]anthracene 2. acenaphthylene 3. acenaphthene 4. naphthalene 5. fluorene 6. anthracene 7. phenanthrene 8. chrysene 9. benz[a]anthracene 10. benzo(k)fluoranthene 11. fluoranthene 12. benzo[a]pyrene 13. pyrene 14. benzo[b]fluoranthene 15. indeno[1,2,3-cd]pyrene 16. benzo[ghi]perylene
449 199 128 271 336 329 230 259 228 279 163 195 125 215 122 115
444 362a 447a 216 303 156 367a 189 345 174 180a 133 374a 228
resolution (R) without R-CD
with R-CD
1.32 7.96 2.81 10.69 1.24 1.43 2.61 1.02 5.54 5.21 10.40 1.02 5.45 1.12 11.70
8.99 4.67 0.76 1.50 2.48 3.72 4.42 2.17 3.73 2.13 2.24 7.30 1.62
a These peaks moved forward with the addition of R-CD to the running buffer. The basic separation buffer consisted of 35 mM sulfobutoxy-β-cyclodextrin and 10 mM methyl-β-cyclodextrin in 50 mM borate, pH 9.2. Numbers before PAHs refer to the peak order in Figure 2c.
Figure 4. Triplicate electropherograms obtained with the same cartridge and adapter. Sample 4000×, 67 cm capillary, 50 mM borate buffer, pH 9.2, containing 35 mM SBβCD, 10 mM MβCD, and 4 mM R-CD.
Figure 3. Electropherograms obtained with 35 mM SBβCD, 10 mM MβCD, 4 mM R-CD in 50 mM borate buffer, pH 9.2. (a) Without microextraction, sample 40×, 57 cm capillary. (b) With microextraction, sample 4000×, 67 cm capillary.
cavity size of 4.9 Å.13 Therefore, one must concede that cyclodextrin-aided capillary electrophoresis is still a trial-and-error procedure, and the interaction mechanism between cyclodextrins and PAHs is yet to be deciphered.
Comparing parts a and c of Figure 2, 11 of the 16 PAH compounds were observed to maintain the same migration order when R-CD was added to the SBβCD/MβCD-containing buffer, whereas five other components exhibited accelerated migration to emerge closer to the solvent peak (see Table 1). The five PAHs affected by the addition of R-CD also showed improvements in plate numbers, while those of the other PAHs remained essentially unchanged. Adding R-CD clearly unclustered the electropherogram; this effect was quantitatively represented by the resolution factors that are mostly higher than 1.5, with the exception of that corresponding to naphthalene, in which the resolution decreased from 2.81 to 0.76. On the contrary, the resolution factors obtained without R-CD include several values slightly higher than unity (for benz[a]anthracene, fluorene, anthracene, and dibenz[a,h]anthracene), indicating comigration (see Table 1). Application of Solid Phase Microextraction. Solid phase microextraction was clearly effective in concentrating the analytes present in a dilute solution. The electropherogram in Figure 3b was obtained for a solution diluted 100-fold from the sample used to obtain Figure 3a, but the peak heights for all species were reduced about 40-60% only. Apparently, when a longer separation capillary (67 cm overall, 60 cm effective length) was used with the SPME procedure, even better separation was achieved for the analytes exhibiting shorter migration times. In addition, the SPME adapter nearly attained zero-volume connection, as manifested by the preservation of the migration order as well as the resolution power. The most crucial characteristic of an SPME adapter is the space or gap between the two capillaries. Their Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1729
Table 2. Reproducibility of the SPME/CE Procedurea peak area cassette 1 peak
fiber 1
fiber 2
fiber 3
fiber 4
mean ( 95%
cassette 2a mean ( 95%
cassette 3a mean ( 95%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
643 2198
625 2182
635 2274
631 2245
633 ( 12 (1.89) 2225 ( 67 (3.01)
636 ( 23 (3.62) 2199 ( 109 (4.96)
638 ( 12 (1.89) 2178 ( 149 (6.84) 6774 ( 100 (1.47)
1123 5589 7418 3328 2680 2234 2613 2273 1376 1585 5642 2296
1076 5439 7352 3482 2576 2085 2654 2254 1403 1497 5732 2308
1152 5439 7285 3362 2654 2234 2741 2197 1378 1469 5703 2341
1097 5465 7342 3407 2675 2187 2705 2265 1376 1604 5705 2354
1112 ( 52 (4.67) 5483 ( 114 (2.08) 7349 ( 87 (1.18) 3394 ( 106 (3.12) 2646 ( 77 (2.91) 2185 ( 112 (5.12) 2678 ( 89 (3.32) 2247 ( 55 (2.45) 1383 ( 21 (1.52) 1539 ( 105 (6.82) 5695 ( 61 (1.07) 2325 ( 43 (1.85)
1145 ( 15 (1.31) 5601 ( 188 (3.36) 7490 ( 143 (1.91) 3384 ( 172 (5.08) 2416 ( 80 (3.31) 2260 ( 123 (5.44) 2163 ( 183 (3.84) 2414 ( 86 (3.56) 2027 ( 60 (2.96) 1544 ( 75 (4.86) 5819 ( 40 (0.69) 2310 ( 92 (3.98)
6818 ( 117 (1.72) 3518 ( 96 (2.73) 2267 ( 87 (3.84) 2287 ( 129 (5.64) 2125 ( 138 (6.49) 2716 ( 75 (2.76) 1688 ( 87 (5.15) 1599 ( 67 (4.19) 5695 ( 61 (1.07) 2172 ( 128 (5.89)
a Results for cassettes 2 and 3 are the averages of four extraction fibers. The values in parentheses indicates the percentage of error (95% confidence interval). For cassette 3, peaks 3, 4, and 5 are integrated together. PAHs corresponding to the peak numbers are identified in Table 1.
ends must be well polished, and they must be positioned as close as possible inside the heat-shrunk tubing to minimize the dead volume. Previous studies indicated that electrophoretic separation and detection of PAHs equipped with UV detection were limited to sample concentrations in the milligrams per liter (ppm) range, whereas laser-induced fluorescence could detect as low as 0.9 µg/L (ppb) of pyrene,4 The detection limit of each analyte was determined by calculating the concentration that would give an integrated peak area equal to 3 times the standard deviation, obtained with 10 repetitions. In this study, the same procedure was applied to data collected with six repetitions. The most sensitive detection could be achieved with pyrene at about 8 ppb, while the highest limit of detection, that for acenaphthene, was 75 ppb. Since these detection limits were obtained with a straight capillary (50 µm i.d.), one may expect an improvement of 4-5fold in sensitivity by using a bubble cell capillary, i.e., a short section of capillary where the internal diameter has been widened to 250 µm, in the detector.3,4 In view of this, as low as 2 ppb of pyrene could be detected by SPME-coupled, cyclodextrin-modified capillary electrophoresis using UV detection and a bubble cell capillary. This detection limit is now competitive with laserinduced fluorescence, which is more expensive and can be only applied to detect fluorescent compounds. Of course, if higher sensitivity is required for analysis of PAHs or other fluorescent analytes, SPME/cyclodextrin-modified capillary electrophoresis equipped with laser-induced fluorescence could be a detection procedure of choice. Reproducibility of the SPME/CE Procedure. A series of experiments was carried out to assess the reproducibility of the SPME/CE procedure. Although it was observed that the 67 cm capillary provided better resolution (Figure 3), the following experiments were performed with 57 cm capillaries since they are commonly used. Three different cassettes were prepared; each consisted of a separation capillary and an adapter to accept an extraction fiber. Very satisfactory reproducibility with respect to migration time (variation 2 h absorption time showed no change in either peak height or peak area (figure not shown). In four repetitions with the same cassette using four different extraction fibers, the variation in the peak area of each PAH was always 500 ppm. On the other hand, up to 100 ng of some compounds has been reported to absorb on the coated fiber; the technique of SPME coupled to CE may have potential for even further improvement. In addition, the assertion of equilibrium between the aqueous phase and the polymer coating as well as a knowledge of the linear range would have to be established before the technique can be used for routine quantitative analyses. It should be noted that, when SPME was first coupled with GC or HPLC, the interaction between the target analytes and microquantity of solvents was not well studied and understood. Undoubtedly, such a phenomenon must be investigated in the CE/SPME technique. Nevertheless, the results obtained so far would suffice to encourage further investigations. In conclusion, for the first time, SPME has been coupled with CE for sensitive analysis of PAHs by using UV detection. The achieved sensitivity is only slightly lower than that reported in the CE equipped with laser-induced fluorescence. The detection limit achieved in this work, combined with the high separation efficiency of cyclodextrin-modified CE, renders the combined CE/ SPME a simple but powerful approach for analysis of PAHs and other non-fluorescent pollutants in contaminated soils and water. It should be noted that PAHs in contaminated soils can be extracted using CO2 supercritical fluid and the extractants collected in dichloromethane, methanol, or a 50/50 mixture of these two solvents.4 The extract can be analyzed for its PAH content using cyclodextrin-modified CE equipped with fluorescence detection. Work is in progress to apply CE/SPME in connection with CO2 supercritical fluid for the detection and quantitation of PAHs in contaminated soils and water.
Received for review September 25, 1996. February 10, 1997.X
Accepted
AC960986O X
Abstract published in Advance ACS Abstracts, March 15, 1997.
Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1731
