Isolation of Priority Polycyclic Aromatic Hydrocarbons from Natural

mental Analysis 13; Elsevier: New York, 1993; Chapter 12. ..... Samples after Their Analysis by Solid-Phase Immunosorbent Extraction Followed by LC-DA...
0 downloads 0 Views 57KB Size
Anal. Chem. 1998, 70, 4996-5001

Isolation of Priority Polycyclic Aromatic Hydrocarbons from Natural Sediments and Sludge Reference Materials by an Anti-Fluorene Immunosorbent Followed by Liquid Chromatography and Diode Array Detection Sandra Pe´rez,† Imma Ferrer,† Marie-Claire Hennion,‡ and Damia` Barcelo´*,†

Department of Environmental Chemistry, CID-CSIC c/ Jordi Girona, 18-26, 08034 Barcelona, Spain, and Ecole Supe´ rieure de Physique et de Chimie Industrielles de Paris, Laboratoire de Chimie Analytique (URA CNRS 437), 10, rue Vauquelin, 75231 Paris, Cedex 05, France

The selective isolation of PAHs from complex environmental mixtures was accomplished by means of a new methodology based on antigen-antibody interactions. This method consists on the extraction of PAHs from water samples onto an anti-fluorene immunosorbent (IS) followed by liquid chromatography with diode array detection. Environmental sediments and a sludge reference material containing PAHs were analyzed using this methodology in order to validate the performance of the IS for the cleanup procedure of such materials. Sediments were extracted by sonication with dichloromethane/methanol (2:1), and the extracts were brought to a volume of 100 mL of water in order to perform the extraction with the anti-fluorene IS. The reliability of the cleanup achieved by the IS was well demonstrated in the analysis of sediment and sludge complex samples containing the priority PAHs established by the U.S. EPA at concentrations varying from 56 µg/kg to 26 mg/kg. Results were compared to those obtained with conventional cleanup procedures showing a better selectivity for PAHs. The chromatograms presented a clear baseline allowing the determination and quantification of PAHs at the ppb level. Using immunosorbents, extraction, trace enrichment, and cleanup were accomplished in only one step. The trace determination of polycyclic aromatic hydrocarbons (PAHs) in environmental samples has been an ongoing challenge for several years. PAHs are one of the most widely measured group of environmental pollutants. PAHs are widespread environmental contaminants resulting from emissions of a variety of sources, including industrial combustion and discharge of fossil fuels, residential heating, and automobile exhaust. Because of their mutagenic and carcinogenic properties, PAHs have been measured in a variety of environmental matrixes, including air, water, soil (sediment), and tissue samples. PAHs are usually present in environmental samples as extremely † ‡

CID-CSIC. URA CNRS 437.

4996 Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

complex mixtures. These mixtures contain many isomer structures and alkylated isomers which greatly vary in relative concentrations of the individual components and in carcinogenic and/or mutagenic properties.1 Due to their environmental concern, PAHs are included in the U.S. EPA and in the European Union priority lists of pollutants. In Europe, the concentration values for the six target PAHs (fluoranthene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, benzo[ghi]perylene, indeno[1,2,3-cd]pyrene) cannot exceed 0.2 ppb in drinking water, whereas in surface water, maximum levels can go up to 1 ppb, depending on the surface water treatment process. The U.S. EPA list of PAHs contains 10 PAHs more in addition to those listed in the European list. Quality control of analytical procedures with the aim to identify and quantify PAHs in environmental matrixes has been a part of the large program of the National Institute of Standards and Technology in the United States1-5 and of the Standards Measurement and Testing Program (SMT), formally known as Community Bureau of Reference (BCR) of the European Commission.6 The analysis of certified reference materials is a key issue when an analytical method is implemented. When a novel method of analysis or a modified method is proposed, the best way to achieve the target results is by analyzing reference materials. This has been the approach of the present paper. Due to the known difficulties in analyzing PAHs in environmental samples, especially as regards to the cleanup procedures, a new methodology involving highly selective isolation by an anti-fluorene immunosorbent (IS) is proposed. This work follows previous work from (1) Wise, S. A. In Environmental analysis: techniques, applications and quality assurance; Barcelo´, D., Ed.; Techniques and Instrumentation in Environmental Analysis 13; Elsevier: New York, 1993; Chapter 12. (2) Wise, S. A.; Hilpert, L. R.; Rebbert, R. E.; Sander, L. C.; Schantz, M. M.; Chesler, S. N.; May, W. E. Fresenius J. Anal. Chem. 1988, 332, 573-582. (3) Wise, S. A.; Benner, B. A.; Christensen, R. G.; Koster, B. J.; Kurz, J.; Schantz, M. M.; Zeisler, R. Environ. Sci. Technol. 1991, 25, 1695-1704. (4) Rebbert, R. E.; Chesler, S. N.; Guenther, F. R.; Koster, B. J.; Parris, R. M.; Schantz; M. M.; Wise, S. A. Fresenius J. Anal. Chem. 1992, 342, 30-38. (5) Hennion, M.-C.; Pichon, V.; Barcelo´, D. Trends Anal. Chem. 1994, 13, 361372. (6) de Voogt, P.; Hinschberger, J.; Maier, E. A.; Griepink, B.; Muntau, H.; Jacob, J. Fresenius J. Anal. Chem. 1996, 356, 41-48. 10.1021/ac980533e CCC: $15.00

© 1998 American Chemical Society Published on Web 10/24/1998

EXPERIMENTAL SECTION Immunosorbent Columns. Preconcentration of the water samples was carried out either off-line or on-line using experimental cartridges prepacked with 0.5 g of silica and 10 mg of antifluorene antibodies and OSP-2 cartridges prepacked with 80 mg of silica and 2 mg of anti-fluorene antibodies. Polyclonal antibodies immobilized on this adsorbent were supplied by Prof. Le Goffic (ENSCP, Paris, France). Polyclonal antibodies were synthesized against fluorene according to the procedure described in a previous study.8 Other preconcentrations using nonselective sorbents were carried out on precolumns prepacked with C18 silica from Merck (Darmstadt, Germany). Chemicals. HPLC-grade solvents acetonitrile and methanol and LC-grade water were purchased from Merck. PAH standards were obtained from Promochem (Wesel, Germany) and deuterated PAH standards from Cambridge Isotopes (Cambridge, U.K.). Dichloromethane, hexane, isooctane, tetrahydrofuran, sodium phosphate, sodium chloride, and azide were obtained from Merck. A stock solution of 10 µg/mL was used to spike LC-grade water and groundwater at the microgram per liter level for preconcentration through the cartridges and further determination of recoveries and construction of the calibration graphs. The phosphate-buffered solution (PBS) consisted of a 0.01 M sodium phosphate buffer containing 0.15 M NaCl (pH 7.4) and 0.2% azide. Reference Materials. (a) EQUATE Project. The EQUATE is an European Union project that has the objective of building institutional capacities in Central and Eastern European countries and implementing in these countries an infrastructure that supports and secures the quality of measurements for national and international water management. Within EQUATE, a program of interlaboratory studies is an important feature. In this respect, the purpose is to provide the participating laboratories tools for diagnosis and improvement. Provision of reference materials and calibrants to the participants of EQUATE supports this process of improvement. In this way, reference river sediment

samples were prepared by EQUATE for conducting interlaboratory studies. Sediment samples were lyophilized before the application of the methodology. (b) BCR Program. The aim of this European program is to certify reference materials to control the quality of the determination of PAHs in various matrixes. In this sense, six priority PAHs were selected as target compounds for certification. The collection, preparation, and homogeneity were performed by the laboratory coordinating the project, including the subsequent interlaboratory studies for certification. The sludge was collected in 1991 in an area with former industrial wood treatment activity leading to soil contamination by creosote oil. Subsequently, the soil was transported to IRMM, Geel (B), where it was dried at 80 °C during 5 days and sieved through a stainless steel sieve of 1-mm mesh size to eliminate stones. Next, the soil (77 kg) was homogenized in a Turbula mixer and subsequently in a jet mill. Particles larger than 125 µm were thus removed. The soil was finally dried and stabilized by heating at 105 °C for 3 h. The remaining material was placed in brown glass bottles under argon, containing ∼40 g of each material. Chromatographic Conditions. LC-DAD analyses were performed with a Waters 600-MS solvent delivery unit with a 20-µL injection loop and a Waters 996 photodiode array detector (Waters, Millipore, MA). Quantification was carried out with UV detection at 254 nm for all the compounds under study. The analytical column used was a 15 cm × 4.6 mm i.d. packed with 3-µm octylsilica gel from Shandon (Cheshire, England). The mobile phases used for the elution of the analytes were acetonitrile and water. The gradient elution was performed as follows: from 50% A (acetonitrile) and 50% B (LC-grade water) kept isocratic for 5 min, to 100% A and 0% B in 20 min, and then isocratic for 5 min more at 1 mL/min. GC/MS analyses were performed with a Carlo Erba GC8000 Series system coupled to a mass spectrometer (Fisons MD800). A 30-m HP-5 column (5% phenyl methyl silicone; 0.25 mm i.d. × 0.25 µm film thickness) was used. The oven temperature program was from 90 to 120 °C at 10 °C/min, from 120 to 320 °C at 4 °C/ min, and then held at 320 °C for 10 min; injector and transfer line temperatures were 280 and 300 °C, respectively. Helium was the carrier gas (50 cm/s). Data were acquired in the electron impact mode (EI) with an electron energy of 70 eV, and the operation was in selected ion monitoring (SIM). The injector was set in the splitless mode (1 µL injected), the split valve being closed for 48 s. Deuterated internal standards for quantification were pyrened10 and perylene-d12. Sample Preparation. (a) Sediment Extraction with Alumina Cleanup. Sediment and sludge samples, previously lyophilized, were extracted by sonication for 20 min with a mixture of dichloromethane/methanol (2:1) three times. The extracts were separated from the sediment by means of a centrifugation step. The three extracts were combined, evaporated to dryness, and redissolved in a mixture of hexane/dichloromethane (19:1). Afterward, the extracts were purified following a cleanup procedure with an alumina column.9 Recoveries of the PAHs obtained by this method and using two internal standards, anthracene-d18 and benzo[ghi]perylene-d12, varied between 70 and 90%. The

(7) Ferrer, I.; Hennion, M.-C.; Barcelo´, D. Anal. Chem. 1997, 69, 4508-4514. (8) Bouzige, M.; Pichon, V.; Hennion, M.-C. J. Chromatogr., in press.

(9) Ferna´ndez, P.; Vilanova, R.; Grimalt, J. O. Polycyclic Aromat. Compd. 1996, 9, 121-128.

our group on the use of immunosorbent columns for trace enrichment of pesticides in water and sediment samples.7 The objectives of the present work were as follows: (i) to develop an analytical method based on an anti-fluorene immunosorbent column for the selective isolation of 13 priority PAHs from EPA using certified sediment and sludge reference materials and (ii) the quantitation of the different isolated PAHs and further comparison of the data with conventional extraction and cleanup procedures and with the values reported in the data sheets of the reference materials given by the organizations. The purpose of the method developed in this paper is to show the advantage of a high selective isolation of PAHs from complex environmental matrixes and the excellent possibilities of the immunosorbent approach to eliminate matrix interferences from extremely dirty river sediments and sludges. To our knowledge no analytical methodology using immunosorbents coupled to liquid chromatography with diode array detection (LC-DAD) has been applied to the determination of PAHs in certified sediment materials.

Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

4997

elution step consisted of washing the column with 5 mL of hexane/dichloromethane (19:1), which removed the most hydrophobic impurities of the sample. Next, the column was eluted with 10 mL of hexane/dichloromethane (1:2), which removed the PAHs. This second fraction was preconcentrated in a rotary evaporator to 0.5 mL, carefully evaporated to dryness with a gentle stream of nitrogen, and brought to a volume of 500 µL with isooctane. Two extracts were analyzed by GC/MS and one by LC-DAD. (b) Sediment Extraction with IS Cleanup. Sediment and sludge samples, previously lyophilized, were extracted by sonication for 1 h with a mixture of dichloromethane/methanol (2:1). The extracts obtained were preconcentrated in a rotary evaporator to 2 mL, carefully evaporated to dryness with a gentle stream of nitrogen, and brought up to a volume of 500 µL with acetonitrile. Afterward, an aliquot of this extract was added to a volume of 100 mL of groundwater in order to get the sample in an aqueous phase and be able to perform the preconcentration through the immunosorbent. So that, the sediment samples could be treated as water samples and follow the usual procedure established in the next section. (c) Water Samples. Water samples were filtered through a 0.45-µm filter (Millipore, Bedford, MA) before use. Preconcentration of the samples was performed either off-line or on-line with automated sample preparation systems such as (ASPEC) XL and OSP-2, respectively. The (ASPEC) XL system, fitted with an external 306 LC pump for the dispensing of samples through the immunosorbent cartridge and with a 817 switching valve for the selection of samples, was a gift from Gilson (Villiviers-le-Bel, France). The second automated SPE device used (OSP-2, Merck) was connected on-line with the gradient pumps. A LiChroGraph model L-600A intelligent pump (Merck-Hitachi) was used to deliver the solvents to condition the precolumns and the water containing the PAHs. The general scheme of the system was described previously.10 Using the off-line procedure, the first step of the solid-phase extraction consisted of conditioning the immunosorbent (0.5 g of bonded silica) with 6 mL of PBS and then with 6 mL of LC-grade water. Afterward, 10 mL of the sample was percolated through the immunosorbent at a flow rate of 1 mL/min. The sample volume of 10 mL was chosen according to the results obtained in a previous work.8 Due to the high hydrophobicity of the PAHs, the addition of 10% acetonitrile to the water samples before their extraction was necessary in order to avoid their adsorption to the flasks.11 The compounds trapped on the immunosorbent were eluted with 4 mL of a mixture containing 70% acetonitrile and 30% LC-grade water. Since PAHs are a very volatile class of compounds, the evaporation from this mixture containing acetonitrile and water would cause loss of the most volatile ones. For this reason a second solid-phase extraction step using a C18 precolumn and a more volatile solvent such as tetrahydrofuran (THF) as the elution solvent was necessary.12 In this case, the precolumn was conditioned with 6 mL of THF, 6 mL of methanol, and 6 mL of (10) Chiron, S.; Papilloud, S.; Haerdi, W.; Barcelo´, D. Anal. Chem. 1995, 67, 1637-1643. (11) Vera-Avila, L. E.; Covarrubias, R. Int. J. Environ. Anal. Chem. 1994, 56, 33. (12) Pichon, V.; Aulard-Macler, E.; Oubihi, H.; Sassiat, P.; Hennion, M.-C.; Caude, M. Chromatographia 1997, 46, 529-536.

4998 Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

Figure 1. LC-DAD chromatograms corresponding to the analysis of a reference sediment material extract after a cleanup step using (a) an alumina column and (b) an immunosorbent column. Peaks: (1) phenanthrene, (2) anthracene, (3) fluoranthene, (4) pyrene, (5) benz[a]anthracene, (6) chrysene, (7) benzo[b]fluoranthene, (8) benzo[k]fluoranthene, (9) benzo[a]pyrene, and (10) benzo[ghi]perylene; *, interference caused by THF.

acetonitrile/water (7:3) followed by percolation of the 4 mL, extracted from the IS, containing the PAHs. Finally, the elution step was performed with 6 mL of THF. The final extracts containing the PAHs in tetrahydrofuran were evaporated with a stream of nitrogen to 500 µL. Finally, a volume of 20 µL was injected into the LC-DAD system. Using the on-line procedure, the first step of the solid-phase extraction consisted of conditioning the immunosorbent with 6 mL of PBS and then with 6 mL of LC-grade water. Afterward, 10 mL of the water sample was percolated through the immunosorbent at a flow rate of 1 mL/min. The compounds trapped on the immunosorbent were eluted with the chromatographic mobile phase by switching the valve into the elute position. For recovery studies, 10 mL of LC-grade water and groundwater sample spiked at 50 and 1 µg/L for off-line and on-line SPE procedures, respectively, were percolated through the immunosorbent. This experiment was performed in triplicate for all the PAHs studied and following the off-line and on-line methodologies in order to compare them. The calibration curves were obtained by percolating 10 mL of groundwater sample spiked in the trace level range of 10-50 µg/L in order to have the same matrix as in the environmental water samples. When the immunosorbent was not in use, it was stored at 4 °C in a solution of PBS after a washing step using 70% methanol and 30% water (5 mL). RESULTS AND DISCUSSION Analytical Performance. Figure 1 shows the comparison between the LC-DAD chromatograms corresponding to the analysis of a reference sediment material extract after a cleanup

Figure 2. LC-DAD chromatograms obtained after the on-line preconcentration of 10 mL of a groundwater sample spiked at 1 µg/L with the mixture of PAHs through (a) a C18 cartridge and (b) an antifluorene immunosorbent. Peaks: (1) fluorene, (2) phenanthrene, (3) anthracene, (4) fluoranthene, (5) pyrene, (6) benz[a]anthracene, (7) chrysene, (8) benzo[b]fluoranthene, (9) benzo[k]fluoranthene, and (10) benzo[a]pyrene.

step using an alumina column and an immunosorbent column, respectively. As can be seen in the second chromatogram (b), the interferences corresponding to the matrix of the sediment sample are fewer than in the first one (a), thus indicating a higher selectivity of the anti-fluorene immunosorbent for the PAHs as compared to conventional cleanup. This proves that the use of immunosorbents is a more reliable technique than the use of alumina for the cleanup of complex environmental samples such as sediments and sludges. As is observed in Figure 1, the cleanup with alumina does not eliminate the extraction of other related compounds from the matrix. Moreover, cleanup steps involving the use of alumina, silica, or Florisil materials are usually cumbersome procedures since the previous elimination of impurities present in these kind of sorbents before their application for the cleanup is needed. Furthermore, the conventional extraction and cleanup cannot be automated as in the case of a methodology using immunosorbents, where less manipulation of the sample is obtained and more reproducible results are achieved. Figure 2 shows the comparison between the LC-DAD chromatograms corresponding to the analysis of 10 mL of a ground water sample spiked at 1 µg/L with the mixture of PAHs after their on-line preconcentration through a C18 cartridge (a) and an anti-fluorene (b) immunosorbent, respectively. As can be observed in (b), the peak corresponding to the matrix of the water is lower than in (a), thus indicating that the matrix of the water

is retained in the C18 cartridge whereas it is not retained in the immunosorbent. Moreover, as is observed in this figure, the interferences encountered with the analysis using immunosorbents are much lower than those obtained with conventional cartridges. This allows a better quantification of the compounds under study since no interference peaks are obtained in the chromatogram and a clear baseline is achieved. A similar behavior was observed when immunosorbents were used for the trace analysis of pesticides in environmental samples.7,13-16 This fact indicates that the use of immunosorbents for the selective solidphase extraction of PAHs from environmental samples is an advantageous technique as compared with the conventional ones that use C18 or polymeric cartridges. Recoveries. In previous studies13 it was shown that a strong affinity for the antigen compound is obtained in immunosorbent columns. Nevertheless, after a rather long period of immunization, the affinity for compounds other than the antigen is also achieved due to the similarity in the chemical structures of compounds between the same family. In the case of an antifluorene immunosorbent, it is expected that antibodies present a certain affinity for other related PAHs than fluorene. This fact was demonstrated in a previous study where the selective isolation of six priority PAHs was accomplished using an anti-fluorene immunosorbent.8 The recoveries of extraction of several PAHs from LC-grade water and groundwater on the anti-fluorene immunosorbent using either an off-line or an on-line methodology are presented in Table 1. These recoveries were obtained after the extraction of 10 mL of a water sample spiked at 50 and 1 µg/L with a mixture of 13 PAHs followed by LC with DAD detection. Recoveries ranged from 7 to 56% for all the compounds studied, indicating a certain affinity of the anti-fluorene immunosorbent for compounds other than the antigen. Incomplete recoveries may be due either to the low affinity presented by the antibodies toward the analytes or to overloading of the capacity of the immunosorbent when related compounds are competing for the same binding sites. In the first case, the recovery is limited by the retention factor of the analyte by the immunosorbent, whereas in the second case, a decrease in the competition process will represent an improvement in the recovery. In our case, when using an off-line methodology, low recoveries of extraction are obtained due probably to an overloading of the immunosorbent caused by the high concentration level present in the water samples. However, such a concentration level was chosen in order to detect appreciable amounts of analytes after the extraction and further evaporation of the extract, taking into account that only an aliquot of the final extract is injected into the chromatograph when an off-line methodology is used. We should also indicate that the methodology developed in this paper included 13 PAHs in order to be able to apply such methodology to real samples. If we had used only six PAHs the recoveries would have been higher, between 35 and 50%, as shown in a previous work.8 (13) Pichon, V.; Chen, L.; Hennion, M.-C.; Daniel, R.; Martel, A.; Le Goffic, F.; Abian, J.; Barcelo´, D. Anal. Chem. 1995, 67, 2451. (14) Pichon, V.; Chen, L.; Durand, N.; Le Goffic, F.; Hennion, M.-C. J. Chromatogr., A 1996, 725, 107-119. (15) Pichon, V.; Rogniaux, H.; Fischer-Durand, N.; Ben Rejeb, S.; Le Goffic, F.; Hennion, M.-C. Chromatographia 1997, 45, 289. (16) Ferrer, I.; Pichon, V.; Hennion, M.-C.; Barcelo´, D. J. Chromatogr., A 1997, 777, 91-98.

Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

4999

Table 1. Extraction Recoveries (%) Obtained after the Percolation of 10 mL of LC-Grade Water and Groundwater Sample Spiked at 50 and 1 µg/L with the Mixture of PAHs through the Anti-Fluorene IS Using Off-Line and On-Line Methodologies, Respectively (n ) 3)a off-line (50 ppb) compound

LC-grade water

groundwater

on-line (1 ppb) groundwater

fluorene phenanthrene anthracene fluoranthene pyrene benz[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene dibenzo[a,h]anthracene benzo[ghi]perylene indeno[1,2,3-cd]pyrene

20 21 21 20 19 10 15 7 13 9 35 21 13

22 26 25 18 16 18 23 7 17 13 25 18 17

16 39 37 50 56 29 30 12 10 11 ndb nd nd

a The relative standard deviation varied between 8 and 25% for offline analysis and between 3 and 15% for on-line analysis. b nd, not detected.

Slightly differences are observed in the retention of the analytes from the two kinds of water as observed in Table 1. This result confirms again the high selectivity of the immunosorbent for the compounds present in any type of water. Selectivity is already achieved by the immunosorbent. The matrix of the water is not retained at all in the immunosorbent, so that the interaction between the matrix and the antibodies is low, thus leading to a high selectivity for the analytes studied. The main phenomenon taking place in the immunosorbent is the competition between the compounds for their binding to the recognition sites of the antibodies.14 However, it is well-known that PAHs are very hydrophobic and tend to stick everywhere. For this reason, nonselective interactions may occur between the sorbent and the analytes leading to a nonspecific retention of these compounds in the immunosorbent. The contribution of the nonselective interactions on the retention of PAHs was already evaluated in a previous work using an anti-atrazine immunosorbent.8 Nevertheless, the retention of PAHs was demonstrated to be better when using specific antigen-antibody interactions for such compounds in the retention mechanism. Table 1 presents also the recoveries of extraction of several PAHs from groundwater on the anti-fluorene immunosorbent using an on-line methodology. In general, an increase in the recoveries is observed as compared with the off-line methodology. This can be attributed to less manipulation of the sample, which avoids loses by evaporation that normally occur with this kind of compound and also to the lower level of concentration used which avoids overloading of the binding sites in the immunosorbent. One advantage of automation in an on-line preconcentration is that more reproducible results are expected, provided that the manipulation of the samples is avoided as compared with an offline methodology.7 Calibration Curves and LODs. Calibration curves were constructed for all the compounds at concentrations ranging from 10 to 50 µg/L (see Table 2). The limits of detection were 5000 Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

Table 2. Calibration Data Obtained with LC-DAD for the Studied PAHs (Spiked from 10 to 50 µg/L) after the Off-Line Preconcentration of 10 mL of Groundwater compound

calibration equationa

R2

LODb (µg/L)

fluorene phenanthrene anthracene fluoranthene benzo[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene dibenzo[a,h]anthracene benzo[ghi]perylene indeno[1,2,3-cd]pyrene

Y ) 4.8 + 20.7x Y ) 18.2 + 45x Y ) 33.9 + 80.9x Y ) 5.5 + 6.2x Y ) 5.3 + 10.2x Y ) 7.6 + 30.6x Y ) 1.9 + 21.2x Y ) 0.4 + 27.1x Y ) 0.02 + 5.96x Y ) 15.5 + 6.1x Y ) 8.6 + 0.39x Y ) 9.4 + 4.6x

0.907 0.894 0.847 0.914 0.895 0.789 0.881 0.996 0.998 0.891 0.899 0.963

7.5 1.5 0.9 15 7.5 3.7 15 30 25 7 22 37

a Least-squares regression equation. b LODs were calculated by using a signal-to-noise ratio of 3 (the ratio between the peak intensity and the noise).

calculated using a signal-to-noise ratio of 3 (the ratio between the peak intensity and the noise). Low detection limits in the ppb level can be obtained due to the high selectivity achieved by the immunosorbents and the high sensitivity encountered by DAD detection. The interaction of the matrix of water and the antibodies is low, thus leading to a high selectivity of the immunosorbent for the analytes studied. The limits of detection depend on the recovery of extraction achieved by the immunosorbent and, therefore, depend on the affinity developed by the immobilized antibodies. The higher the affinity is for a compound the lower is the detection limit. Validation of the Method. This study was mainly focused on the selective extraction of PAHs from complex environmental samples such as sediments and sludges. A way to validate the performance of an IS is to apply it to the extraction of analytes from real samples and to study the effect of the matrix in such complex mixtures. Quality control of analytical procedures aimed at determining PAHs in environmental samples requires the availability not only of PAH calibration standards of high purity but also of matrix materials in order to verify all steps of the analytical procedure. For this purpose one certified sludge material from BCR (Community Bureau of Reference, no. 524, code 005) and two reference sediment materials from EQUATE were extracted and analyzed by following the methodology developed in this work. The aim was to apply the selective isolation presented by the immunosorbents to the extraction of PAHs in such complex samples and compare it with the results obtained with conventional cleanups using alumina. In Table 3, the results obtained using the methodology developed in this work for the three reference materials studied, as well as the certified values, are reported. Some compounds were not detected due to the low concentrations at which they were present in the samples. Pyrene was not detected due to a coelution with an interference peak obtained when the off-line methodology was used. On the other hand, phenanthrene, anthracene, fluoranthene, and benz[a]anthracene were well detected and the concentration value did not differ significantly from the certified value. One should take into account that this kind of sediment and sludges are a very complex class of samples

Table 3. Mean Concentration (N ) 3) and Target Values of the PAHs Analyzed in Reference Sediment and Sludge Samples after Their Analysis by Solid-Phase Immunosorbent Extraction Followed by LC-DADa BCR mg/kg compound phenanthrene anthracene fluoranthene pyrene benzo[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene dibenzo[a,h]anthracene benzo[ghi]perylene indeno[1,2,3-cd]pyrene

mean value

nqb 25.8 23.7 13.2 ndc nd

EQUATE 96049 µg/kg

EQUATE 96050 µg/kg

certified value

mean value

certified value

mean value

certified value

ncc nc nc 173 (6) 22.5 (8) nc 13.5 (12) 6.2 (10) 8.6 (6) nc nc 5.1 (8)

312 56 235 nq 789 394 596 332 nd nd 151 nd

201 (8) 56 (10) 326 (11) nc nc 175 (9) nc 138 (8) nc 34 (11) 107 (14) 133 (14)

569 342 831 n.q. 622 392 319 168 nd nd 157 276

537 (12) 342 (9) 953 (13) 837 (11) 556 (11) 549 (12) nc nc nc nc nc 407 (12)

a The relative standard deviation varied between 10 and 35%. The RSD for the certified value given by the organization is also indicated. b nq, not quantified due to coelution. c nd, not detected. d nc, not certified by the organization.

containing high concentrations of different contaminants and that the matrix is very complex. The differences in concentrations for all the PAHs in real samples increase the competition processes for the binding sites of the immunosorbent, and as a consequence, recoveries can be very different from those calculated at equal concentrations of PAHs. The higher percentage of error is found in the most apolar PAHs. This can be explained by the fact that the immunosorbent contains antibodies selective for fluorene, which is a relative small molecule with a lower molecular weight as compared to the higher molecular weight PAHs. The small PAHs are expected to be better recognized by the anti-fluorene immunosorbent than higher molecular weight PAHs. As can be seen in Figure 1, the chromatograms obtained from the sediment extracts analyzed after their preconcentration through the anti-fluorene immunosorbent presented a clear baseline as compared with those analyzed after the cleanup step using alumina, so that the compounds could be easily identified and detected. Table 4 shows the comparison between the results obtained after the analysis of a certified reference sediment from EQUATE using two different cleanup procedures: alumina and immunosorbent. These results demonstrate that for some compounds these kinds of immunosorbents are useful for the selective extraction of PAHs from environmental samples although at high levels of concentration important percentages of error are obtained. Generally, IS give an overestimation of the results as compared to conventional alumina cleanup, which can also be attributed to an insufficient cleaning between cleanup steps. It has been demonstrated that most of the compounds present in complex sediment and sludge samples could be identified and quantified easily using immunosorbents. So that the selective extraction of PAHs from complex environmental samples is possible using an anti-fluorene immunosorbent as shown in this study.

Table 4. Comparison of the Results (µg/kg) Obtained with Two Different Cleanup Steps for the Analysis of a Reference Sediment Sample from EQUATE (96049)a

compound

cleanup with aluminab

cleanup with ISc

certified value

phenanthrene anthracene fluoranthene chrysene benzo[k]fluoranthene benzo[ghi]perylene

196 60 386 250 249 138

312 56 235 394 332 151

201 (8) 56 (10) 326 (11) 175 (9) 138 (8) 107 (14)

aThe relative standard deviation varied between 10 and 35%. Analysis were performed by GC/MS. c Analysis were performed by LC-DAD.

b

Overall, the reported methodology represents a new approach to sample handling in environmental analysis. The immunosorbent columns can be used at least 100 times for analysis of PAHs since there is always a washing step at the end of the process of extraction which removes the traces of contaminants not eluted from them. Improvements in regard to the percent recovery values of the PAHs are needed. In this respect, batteries of two anti-PAHs immunosorbents will be tested in the future in order to improve the development of an immunosorbent methodology for the whole range of PAHs. ACKNOWLEDGMENT This work has been supported by the Commission of the European Communities, Environment & Climate Program 199498 (ENV4-CT95-0016) and CICYT (AMB96-1600-CE). Received for review May 14, 1998. Accepted August 25, 1998. AC980533E

Analytical Chemistry, Vol. 70, No. 23, December 1, 1998

5001