Anal. Chem. 1999, 71, 2171-2178
Spherical Clay Conglomerates: A Novel Stationary Phase for Solid-Phase Extraction and “Reversed-Phase” Liquid Chromatography Thomas D. Bucheli, Stephan R. Mu 1 ller,* Patrick Reichmuth, Stefan B. Haderlein, and Rene´ P. Schwarzenbach
Swiss Federal Institute for Environmental Science and Technology (EAWAG) and Swiss Federal Institute of Technology (ETH), CH-8600 Du¨ bendorf, Switzerland
A new solid phase is presented to be used for the solidphase extraction (SPE) of organic compounds from aqueous solutions and as a stationary phase for the separation of organic compounds in “reversed-phase” HPLC. The material consists of spherical clay conglomerates (SCCs) in the size ranges of 2-5, 5-10, and 10-20 µm. SCCs are especially well suited for the extraction and separation of aromatic compounds with electron-withdrawing substituents, because of the formation of specific electron donor-acceptor (EDA) complexes of such compounds with natural clay minerals. A series of nitroaromatic compounds (NACs), e.g., nitrophenols, and nitrotoluenes, served as probe substances for the characterization of the SPE with SCCs online coupled to a C18-HPLC-DAD system. Breakthrough volumes were > 1 L and method detection limits (MDLs) < 100 ng/L for compounds with moderate to high affinity towards clay minerals. The performance of the material is hardly affected by matrix effects and because of its excellent physical properties, i.e., regenerability and pressure-resistance, it meets the requirements for fully automated routine trace analysis of several primary pollutants, such as 6-methyl-2,4dinitrophenol (DNOC) or 2,4,6-trinitrotoluene (TNT), in various natural waters. Offline SPE with SCCs was superior or equivalent to commercial SPE products for analysis of such compounds. Finally, SCCs are shown to be well suited as a stationary phase in reversed-phase HPLC. This opens a wide range of applications, e.g., as an easy and fast separation technique that is orthogonal to C18 reversed-phase HPLC. The low concentrations of numerous organic pollutants in natural waters have inspired the development of suitable solidphase extraction (SPE) and HPLC materials for many years. Currently, these materials cover a wide range of polarity, e.g., from hydrophobic to ionic; however, the complex matrix present in natural waters and enriched samples is still the most challenging problem to overcome in order to precisely and accurately quantify organic compounds. Therefore, target-compound-specific SPE materials and stationary phases “orthogonal” to reversed-phase HPLC would improve analytical chemists’ tools tremendously. * Corresponding author: (Fax) +41 1 823 54 71, (e-mail)
[email protected]. 10.1021/ac981133u CCC: $18.00 Published on Web 05/01/1999
© 1999 American Chemical Society
Recent findings revealed that planar aromatic compounds with strongly electron withdrawing substituents, e.g., -NO2 and -CN, form strong complexes with clay mineral surfaces.1 Such compounds include a variety of environmental priority pollutants such as dinitrophenol herbicides, nitrated polycyclic aromatic compounds, and nitroaromatic explosives or nitro musks. Coplanar electron donor-acceptor (EDA) complex formation between the electron deficient π-system of the aromatic solute (electron acceptors) and oxygens present at siloxane surface(s) (electron donors) has been identified as the governing adsorption mechanism. Among the various groups of phyllosilicates, some smectite clays such as montmorillonite exhibit both high affinity and capacity for such substances.1 In this work, a solid phase for SPE and HPLC has been produced and evaluated that exhibits an extremely high selectivity for π-acceptors compared with existing SPE materials. It meets all the important criteria of SPE and HPLC, such as physical and chemical stability in aqueous systems and towards organic solvents, high permeability, and mass-transfer kinetics as well as efficient regeneration. These demands are not met by clay minerals in their natural form. Smectite clay minerals are highly disperse materials (typical particle sizes ,1 µm) and are subject to swelling and shrinking processes in aqueous environments. Basically, two different techniques for modification of natural clay minerals so that they may serve as a solid phase have been used: The first one can be referred to as the “clay-coating” technique, where clay minerals are attached to a solid phase, such as alumina2 or silica gel.3,4 The second one comprises dry spraying of clay-mineral suspensions to produce (quasi-) spherical clay conglomerates (SCCs).5,6 Such material was used in HPLC for separation of metal complexes4 and amino acids.7 Furthermore, its applications in supercritical-fluid chromatography6 and gas chromatography8 were described. SCCs have also been used for normal-phase HPLC separation of various aromatic compounds.5,6,9 However, the (1) Weissmahr, K. W.; Haderlein, S. B.; Schwarzenbach, R. P. Soil Sci. Soc. Am. J. 1998, 62, 269-378. (2) Szecsody, J. E.; Streile, G. P.; Pavalko, W. J. Environ. Sci. Technol. 1993, 27, 356-365. (3) Ko ¨rdel, W.; Stutte, J.; Kotthoff, G. Sci. Total Environ. 1995, 162, 119-125. (4) Nakamura, Y.; Yamagishi, A.; Iwanoto, T. Clay Sci. 1990, 8, 17-23. (5) Tsvetkov, F.; Mingelgrin, U.; Gal, M. J. Therm. Anal. 1994, 42, 113-129. (6) Jinno, K.; Mae, H.; Yamaguchi, M.; Ohtsu, Y. Chromatographia 1991, 31, 239-242. (7) Tsvetkov, F.; Mingelgrin, U. Clays Clay Miner. 1987, 35, 391-399. (8) Baksh, M. S. A.; Yang, R. T. Sep. Sci. Technol. 1991, 26, 1377-1394.
Analytical Chemistry, Vol. 71, No. 11, June 1, 1999 2171
Figure 1. Chemical structures, names, abbreviations, and sorption constants, Kd (in [L/kg]), given for K-montmorillonite (for details see Haderlein et al.11) of the compounds investigated (n.a., no Kd available).
systems described in the literature were not designed and operated to exploit EDA interactions as the predominant retention mechanism of analytes, and their chromatographic performance was not satisfactory in terms of separation power. The SCCs used in our study were further modified to enhance their affinity for π-acceptors, to improve their mechanical and chemical stability in aqueous systems, and to reduce their sensitivity toward matrix constituents typically present in natural waters such as electrolytes and dissolved humic or fulvic matter. The performance of the newly developed SPE material was evaluated using a set of model π-acceptors covering a wide range of complexation constants for phyllosilicates (for compound structures and complexation constants, see Figure 1). Nitroaromatic compounds (NACs) were selected as model analytes because they include a number of priority pollutants such as the herbicide 2,4-dinitro-6-methylphenol (DNOC) or the explosive 2,4,6-trinitrotoluene (TNT) and because their EDA complexes with clay minerals are well-documented.10,11 On-line SPE (with SCCs)HPLC (with C18) breakthrough volumes and method-detection limits (MDLs) of the analytes were determined, and the effects (9) Tsvetkov, F.; Heller-Kallai, L.; Mingelgrin, U. Clays Clay Miner. 1993, 41, 527-536.
2172 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999
of matrix constituents present in aqueous samples of various origins on analyte recoveries were studied. The potential and the limitations of our SPE method for selective enrichment of trace amounts of π-acceptor analytes from aqueous samples was further evaluated using other environmentally relevant compounds such as nitro musks, benzonitriles, nitrodiazobenzenes, and nitroanilines. Furthermore, the suitability of the setup for offline SPE was demonstrated and compared with that of commercially available offline SPE materials. The SCCs were also evaluated as a solid phase in HPLC for analytical separation of aromatic compounds with π-acceptor properties. Moreover, an application of the time-efficient, accurate, and precise online (SCCs)SPE-HPLC method is illustrated by the atmospheric-washout study of nitrophenols (NPs) during rain events. EXPERIMENTAL SECTION Materials. 2-Nitrophenol (2-NP), 2,4-dinitrophenol (2,4-DNP), 1,3-dinitrobenzene (1,3-DNB), 4-methylphenol (p-cresol), thiourea, (10) Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci. Technol. 1993, 27, 316-326. (11) Haderlein, S. B.; Weissmahr, K. W.; Schwarzenbach, R. P. Environ. Sci. Technol. 1996, 30, 612-622.
and alizarin yellow R were from Fluka AG, (Buchs, Switzerland). 4-Amino-2,6-dinitrotoluene (4-A-2,6-DNT) and 2,6-diamino-4-nitrotoluene (2,6-DA-4-NT) were purchased from Promochem (Wesel, Germany). 6-Methyl-2,4-dinitrophenol (DNOC), 6-sec-buthyl-2,4dinitrophenol (dinoseb), pendimethalin, chlorothalonil, and dichlobenil were from Riedel-de Hae¨n (Seelze, Germany). 2,4,6Trinitrotoluene (TNT) was provided from Ems Chemie (Dottikon, Switzerland). Musk ketone, musk tibetene, and musk xylene were a kind donation from Givaudan-Roure Research AG (Du¨bendorf, Switzerland). The purchased 2,4-DNP was purified by recrystallization in ethyl acetate. All other compounds were used as received. Ethyl acetate (purity for pesticide-residue analysis) was from Burdick & Jackson (Muskegon, MI). Sodium silicate solution (14% NaOH, 27% SiO2), acetonitrile (ACN), HCl (32%), LiCl, NaCl, NaOH, CsCl, and montmorillonite K10 were purchased from Fluka AG (Buchs, Switzerland). The clay mineral is sold in a protonated form with a cation-exchange capacity (CEC) of 30-40 mVal/100 g. Nitrogen (99.995%) was from Carbagas (Ru¨mlang, Switzerland). Deionized water was further purified with a Nanopure waterpurification device (NANOpure 4, Skan, Basle, Switzerland). Production of Spherical Clay Conglomerates (SCCs). Exchangeable cations present at siloxane surfaces to compensate their negative charge may hinder EDA complex formation in aqueous solutions, depending on the size of their hydration sphere. Conversely, EDA complex formation of a given clay mineral can be enhanced by substitution of exchangeable cations by weakly hydrated Cs+, e.g., by addition of a Cs-salt as a background electrolyte to the aqueous samples. Alternatively, the amount of exchangeable cations at clay minerals may be permanently reduced by thermal treatment in the presence of Li+. At elevated temperatures, Li+ ions dehydrate and diffuse into the mineral structure to compensate the net negative charge of clays within the crystal lattice, a process which further facilitates EDA complex formation. Both approaches were used (see below). Montmorillonite K10 (20 g/L) was suspended in a 1.6 mM NaOH solution (to obtain a pH ∼6) and set aside for sedimentation of the larger particles. After one week, the suspension was transferred, and NaCl was added (0.01 M) for coagulation and sedimentation of the remaining smaller particles. Sodium was chosen as it is cheaper and more easily exchangeable than other cations. Twenty-four h later, the supernatant was discharged and the sediment collected. The average size of the resulting fine fraction was 90% up to the investigated sample volume of 1 L (see Figure 4). 4-A-2,6-DNT, however, showed decreasing recoveries at higher volumes (61 ( 9% at 1000 mL), and 2-NP and 2,6-DA-4-NT revealed a drastic reduction in recovery, already, at a volume of 200 mL (see Figure 4). Hence, SPE with SCCs of analytes with affinities for clays equivalent or less than that of 2-NP (Kd ) 45 L/kg, Figure 1) do not quantitatively adsorb and therefore do not interfere with the target compounds. p-Cresol illustrates the high selectivity and the completely different sorption mechanism of SCCs: p-Cresol has a similar Kow to NACs (log Analytical Chemistry, Vol. 71, No. 11, June 1, 1999
2175
Kow ) 1.9), but a very different tendency for EDA complex formation at clays (e.g., Kd ) 0.9 L/kg on Cs-kaolinite10). Consequently, p-cresol was not extractable with SCCs at all! The illustrated performance (in terms of recovery and precision) was obtained with all online setups in use, including Cs-, or Lihomoionized materials and forward and backflush elution. Further, a quite sophisticated application of NPs and NTs enrichment at pH 2 followed by sequential elution of, first, the NPs by a pH change from acidic to neutral only (deprotonation), and second, the NTs with an ACN gradient illustrates the specific sorption mechanism as well as the performance and stability of the SCCs over a wide range of chemical parameters. These findings prove that the treatment during production of the SCCs (e.g., heating) did not have any discernible effect on the specific sorption characteristics observed with native clay minerals, but improved the chemical and physical stability ideally. Our target compounds, the primary pollutants DNOC and TNT, as well as certain metabolites and byproducts (e.g., 2,4-DNP, 1,3-DNB, and 4-A-2,6-DNT) can be quantified in Nanopure water in concentrations as low as a few ng/L (for MDLs, see Table 1). In general, the breakthrough volumes and MDLs may be anticipated for any substance for which the Kd on natural clay minerals is known. Roughly, any compound exhibiting a Kd > 50 L/kg on K-montmorillonite11 or Cs-kaolinite10 has the potential to be quantitatively and selectively extracted with the SCCs and quantified at concentrations below 100 ng/L (other examples are given below). Matrix Effects, MDLs, and Precision of the Method for Natural-Water Samples. For the determination of recoveries in various natural waters, samples were fortified with 0.1 µM of the analytes, and the sample volume was 100 mL, i.e., well below the breakthrough volume of most compounds. For most of the investigated compounds, the environmental matrix did not cause a significant alteration in recoveries compared with the ones found in Nanopure water (see Tables 1 and 2). MDLs of NPs in 250 mL of rainwater were determined as described in Table 1, and a chromatogram is shown in Figure 5. The resulting values were 24 ng/L for DNOC (n ) 3, 103 ( 1% recovery) and 27 ng/L for 2,4-DNP (n ) 3, 100 ( 1% recovery). Average precision, as determined with multiple determinations of spiked rainwater samples (n ) 3, 10 nmol in 250 mL), was 1.2% for DNOC, and 2.4% for 2,4-DNP. These numbers are also in good agreement with the MDLs determined with Nanopure water (Table 1). Only 2-NP and 2,6-DA-4-NT, i.e., the analytes with only moderate affinity toward clay mineral surfaces, suffered a reduction in recovery in waters with elevated levels of DOC or ionic strength (Table 2). Extremely high DOC concentrations and high ionic strength, as present in samples from the hypolimnion of a lake receiving effluents of a superfund site,15 also reduced the recovery of 2,4DNP. Conclusively, recoveries of most analytes were not negatively affected in most natural-water samples. For example, priority pollutants such as DNOC or TNT revealed no decrease in recoveries in aqueous samples containing DOC in concentrations as high as 100 mg/L. The MDLs for substances with high breakthrough volumes (>1 L), e.g., DNOC and TNT, could be further reduced by increasing the sample volume, although the extraction time would also increase in the presented setup. However, embedding the SCCs in, e.g., 3M’s or Supelco’s SPE 2176 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999
Figure 6. Offline SPE of 500 mL of contaminated groundwater (Landfill of Riet, Canton Zurich) spiked with 0.5 µM of the investigated NTs and NB on SCCs ((a) Li-homoionized, 5-10 µm and on Porapak RDX b)). The DOC content of this sample was around 20 mg/L. The interference of the environmental matrix was clearly enhanced in chromatogram b compared with a. The smaller 2,6-DA-4-NT peak in chromatogram a illustrates the specific sorption characteristics of SCCs.
disks could considerably shorten the enrichment time. Applicability of SPE with SCCs for the Enrichment of Other Important Pollutants. In addition to the compounds discussed above, some other important pollutants were used to assess the potential of SCCs as an SPE material. An additional nitrophenol pesticide, three nitro musks, a nitroaniline herbicide, a representative of the nitroazocompounds, and two of the benzonitrile pesticides were analyzed by online (SCCs)SPE-HPLC (for structures see Figure 1). The recoveries obtained from Nanopure and various natural waters are given in Table 2 (sample volume ) 100 mL, spike level ) 0.1 µM, n ) 3). Despite the bulky secbutyl substituent of dinoseb (which leads to a reduced Kd as compared with DNOC, Figure 1) and the terbutyl substituents of the nitro musk odorants, high recoveries between 90 and 110% were achieved in Nanopure and different natural waters for these substances. Pendimethalin showed, despite its secondary amine group, good recoveries of 89 and 98% from Nanopure and lake water, respectively. Alizarin yellow R, a large planar aromatic compound consisting of two phenyl rings connected by an azo group, revealed a recovery of 81 ( 2% from Nanopure water and illustrates the potential of the SCCs for enrichment of certain azo dyes. Dichlobenil and chlorothalonil, which are aromatic compounds with chlorine and nitrile substituents that have less
Table 2. Recoveries of Compounds Investigated in Various Natural Waters Using Online SPE-HPLC with Li-Homoionized SCCsa Nanopure water matrix parameters DOC (mg/L) conductivity (µS/cm) NPs DNOC 2,4-DNP Dinosebg 2-NP NTs and NB TNT 1,3-DNB 4-A-2,6-DNT 2,6-DA-4-NT nitro musks musk ketone musk tibetene musk xylene nitroaniline pendimethalin nitro diazobenzene Alizarin yellow R benzonitriles chlorothalonil dichlobenil
rainwater
lakewater Bc
Ab
roof runoff
9 455
100 2150
wastewatere 6
103 ( 1g 100 ( 1g
96 ( 1 95 ( 1
92 103
91 ( 3 68 ( 3
95 ( 1
79 ( 4
95
39 ( 43
101 ( 0 102 ( 0 104 ( 0 39 ( 2 100 ( 2 105 ( 6 107 ( 5
Cd
92 ( 3
MDL (ng/L)
groundwaterf 16 2140
98 ( 0 96 ( 0
99 ( 2 89 ( 1
94 ( 1
9(3
94 ( 2 102 ( 1 101 ( 3
90 ( 1 103 ( 3 109 ( 1
24g 27g 27d
97 ( 2 101 ( 2 99 ( 3 34 ( 5
97 ( 1 104 ( 1 110 ( 2
89 ( 3h
98 ( 2
12d
81 ( 2i 100 ( 2 100 ( 1i
101 ( 0
88 ( 2
100 ( 1
103 ( 0 24d
a No analytes detected in unspiked waters; if not otherwise stated, sample volume 100 mL, solute concentration 0.1 µM, and n ) 3. b Epilimnion of Lake Woburn, MA; n ) 1 for NPs and n ) 2 for NTs and NB. c Hypolimnion of Lake Woburn, MA; n ) 2. d Greifensee (Canton Zurich); sample volume 500 mL, sample concentration nM. e Effluent from a wastewater treatment plant (Opfikon-Glattbrugg, Switzerland). f Landfill-leachatecontaminated groundwater (Riet, Switzerland). g Sample volume 250 mL, solute concentration 0.04 µM. h Solute concentration 10 nM. i n )2.
Table 3. Recoveries Obtained with Offline SPE of NTs and NB with Li-Homoionized SCCs and Comparison with Commerical SPE Materialsa SCCsb
NTs TNT 4-A-2,6-DNT 2,6-DA-4-NT NB 1,3-DNB
Porapak RDX
Chromabond HR-P
Nanopure water
wastewaterc
groundwaterd
Nanopure water
wastewaterc
groundwaterd
Nanopure water
wastewaterc
groundwaterd
92 ( 7 92 ( 7 34 ( 7
93 ( 4 92 ( 4 3(0
94 ( 6 92 ( 6 2(0
88 ( 1 79 ( 1 88 ( 2
90 ( 2 80 ( 2 90 ( 3
92 ( 2 83 ( 2 94 ( 1
80 ( 7 86 ( 5 82 ( 3
76 ( 2 84 ( 1 84 ( 1
83 ( 2 90 ( 1 93 ( 2
90 ( 6
93 ( 3
93 ( 6
89 ( 1
92 ( 1
94 ( 2
85 ( 6
80 ( 4
89 ( 1
a No analytes detected in unspiked waters; sample volume 500 mL, sample concentration 0.5 µM, and n ) 3. b Li-Homoionized, 5-10 µm. Effluent from the wastewater treatment plant in Opfikon-Glattbrugg (Canton Zurich, Switzerland). d Contaminated groundwater from the Landfill of Riet (Canton Zurich).
c
electron-withdrawing strength than the nitro group (see Figure 1), were quantitatively enriched from 100-mL samples (see Table 2). Comparison with Commercial SPE Materials. Figure 6 shows typical chromatograms of extracts obtained from offline SPE with SCCs (Figure 6a) and a commercial product (Figure 6b). Sample volumes were 500 mL, and concentrations were set to 0.5 µM. Chromatograms of environmental samples enriched with SCCs showed much smaller interferences from the matrix present in natural samples as compared to the commercial SPE material (compare the elution profile between 1 and 4 min in Figure 6a,b, which nicely illustrates the selectivity of the SCCs material). Table 3 shows the recoveries obtained with SCCs and commercially available products from Nanopure and natural waters. Most of the investigated compounds revealed persistently high recoveries irrespective of the environmental matrix when extracted
with SCCs. The recovery of 2,6-DA-4-NT, which has a fairly low affinity for clay minerals (see Figure 1), was reduced. Hence, similar to the results described for online SPE-HPLC, the specific affinity of the analytes toward clay minerals determines the performance of SCCs in offline SPE of NTs and NB. Compared with recoveries obtained with SPE using commercially available products, recoveries with SCCs were similar or slightly higher for 4-A-2,6-DNT, 1,3-DNB, and TNT. Only for 2,6-DA-4-NT the commercial products clearly exhibited better recoveries in both Nanopure and environmental waters. SCCs as Stationary Phases for Separation of NACs by Reversed-Phase HPLC. SCCs revealed a huge potential to serve as a stationary phase in HPLC columns. Figure 7 shows the baseline separation of three NTs and one NB (20 µL 10 µM injected) on a 5-cm column packed with Cs-homoionized SCCs of 10-20 µm. Plate numbers of a column packed with Lihomoionized SCCs (particle size 5-10 µm) were ∼6000/m Analytical Chemistry, Vol. 71, No. 11, June 1, 1999
2177
Figure 7. Separation of three NTs and one NB on a SCCs packed column (Cs-homoionized, 10-20 µm). Linear gradient from 100% water (0.02 M CsCl) to 1:1 water (0.02 M CsCl)/ACN within 15 min.
Figure 8. Wet-deposition dynamics of 2,4-DNP during two rain events under different meteorological conditions: cold front from western direction (April 26, 1997) and local thunderstorm (July 18, 1997).
(conservative tracer: thiourea), which is only a factor of 5-10 less than numbers achieved with commercially available HPLC separation columns. A smaller particle size (e.g., 3 µm), a more narrow size distribution, and a specifically adapted slurry packing technique would certainly lead to a comparable performance. The production of such particles in an industrial process should be easy and cost-effective; however, it was beyond the scope of this work. As the retardation of NACs in SCCs separation columns is based on a completely different separation mechanism compared with C18 columns (note that the analyte-retention time correlates with Kd, see Figures 7 and 1), such columns could be used as an alternative (orthogonal separation mechanism) to the conventional reversed-phase columns. Note that, on the basis of the findings of the SPE experiments, we expect no significant adverse effects of DOC on the separation of NACs on SCCs columns. Among other applications, SCCs separation columns might be useful for (low-budget) confirmation of compound identity and/or where separation of certain NACs (e.g., isomers such as 2,6-DA-4-NT and 2,4-DA-6-NT) might pose a problem.
2178 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999
CONCLUSIONS The extraction force of the newly developed and characterized SPE material (SCCs), which was tailored from commercially available clay minerals, is completely different from currently available SPE materials. This leads to its major advantages of high selectivity for organic compounds with electron-withdrawing substituents and suppression of matrix constituents such as humic and fulvic acids. Further, the SCCs are completely regenerable, are pressure-resistant, and can be used in a fully automated enrichment, separation and detection system (online SPE-HPLC). Therefore, the SCCs are perfectly suited for the cost-effective routine analysis of organic pollutants with electron-withdrawing substituents in various kinds of environmental samples. The minimized sample preparation and the high accuracy and precision of the method allowed investigations on the deposition behavior of, e.g., nitrophenols in different rain events as illustrated in Figure 8 (see Figure 5 for an example of a chromatogram). Measured concentrations ranged from almost 4 µg/L at the onset of certain rain events to numbers close to the quantification limit of ∼0.08 µg/L. A detailed discussion of the processes that influence the washout dynamics of NPs and other contaminants is given elsewhere.18 (Offline) SPE with SCCs was superior or equivalent to commercial SPE products for analysis of, e.g., TNT and certain metabolites. A huge potential of SCCs lays, however, in their application in reversed-phase HPLC. As a result of their completely different retention mechanism, they might compete with many C18-HPLC applications or complete them as an orthogonalseparation technique, altogether leading to a better separation of matrix constituents and target compounds. Considering the matrix constituents in human fluids, the SCCs could also be used for the enrichment/purification of some pharmaceutical compounds with electron-withdrawing substituents such as chloramphenicol, nitrazepam, nitrendipine, and many more. ACKNOWLEDGMENT This work is part of the EAWAG Priority Research Program 1993-1997: Sustainable Management of Resources. We are most thankful to K. Weissmahr for all the fruitful discussions on NACs and clay minerals. M. Vogt is acknowledged for his help in the development and application of the presented method. Thanks are due to Y. Weidmann for the electron micrographs. We are grateful to F. Schu¨rch and D. Iseli from Macherey-Nagel for their support with column packing, and to K. Lerch from GivaudanRoure Research AG for the donation of the nitro-musk compounds. We are indebted to A. Alder, M. Berg, R. Knochenmuss, J.-M. Stoll, and M. Suter for reviewing the manuscript. Received for review October 16, 1998. Accepted March 10, 1999. AC981133U (17) Keith, L. H. Environmental Sampling and Analysis, A Practical Guide; Lewis: Boca Raton, FL, 1991. (18) Bucheli, T. D. Swiss Federal Institute of Technology (ETH) no. 12414. Ph.D. Thesis, Zu ¨ rich, 1997.
