Environ. Sci. Technol. 2008, 42, 6415–6423
Analysis and Sorption of Psychoactive Drugs onto Sediment KAROLINE STEIN,† MARIA RAMIL,† GUIDO FINK,† MICHAEL SANDER,‡ AND T H O M A S A . T E R N E S * ,† German Federal Institute of Hydrology (BfG), D-56068 Koblenz, Am Mainzer Tor 1, Germany, and Institute of Biogeochemistry and Pollutant Dynamics (IBP), Swiss Federal Institute of Technology Zurich (ETHZ), Universita¨tstrasse 16, CH-8092 Zurich, Switzerland
Received November 27, 2007. Revised manuscript received March 7, 2008. Accepted March 10, 2008.
An analytical method was developed to analyze eight psychoactive pharmaceuticals—including the antiepileptic carbamazepine, the opiates morphine, codeine, dihydrocodeine, the opiode tramadol, and the tranquilizers diazepam, oxazepam, temazepamsand the antibiotic sulfamethoxazole as well as three metabolites (10,11-dihydrocarbamazepine (DHC), 10,11dihydroxy-10, 11-dihydrocarbamazepine, and N4-acetylsulfamethoxazole) in river sediments. Relative recoveries of all analytes exceeded 97% using either deuterated or 13C15N-labeled surrogate standards. Sorption isotherms of all analytes were constructed at pH 6.5-6.6 on two natural river sediments (Burgen and Dausenau) that differed in organic carbon contents and particle size distributions. Affinities of all analytes were up to an order of magnitude higher for the Dausenau sediment in comparison to the Burgen sediment. Isotherms were well described by the Freundlich model. Sorption of all analytes was linear on the Burgen sediment except for structurally similar carbamazepine (n ) 0.90) and DHC (n ) 0.88). Conversely, most analytes showed pronounced nonlinear sorption to the Dausenau sediment (n ) 0.77-0.92) except for positively charged codeine, dihydrocodeine, and tramadol. Linear sorption of the latter was taken to arise from concentration-independent electrostatic interactions of the organocations with negatively charged surfaces on clay minerals or in the sediment organic matter. Desorption gave rise to hysteresis in 13 out of 16 investigated analyte-sorbent systems. Hysteresis was likely due to slow desorption kinetics beyond 24 h used in the experiment.
Introduction Pharmaceuticals represent an important class of emerging organic micropollutants. More than 100 pharmaceuticals of different therapeutic classes have been detected in rivers and streams. These compounds are introduced to surface waters mainly by discharges of wastewater treatment plants (WWTP) and by runoff from agricultural fields (1–5). Several recent studies report the presence of illicit psychoactive drugs in rivers. Zuccato et al. (6) detected the opium-alkaloid cocaine and its metabolite benzoylecgonine in the Po River at concentrations of 1.2 and 25 ng/L, respectively. Huerta* Corresponding author phone: +49 261-1306 5560; fax: +49 2611306 5363; e-mail:
[email protected]. † German Federal Institute of Hydrology. ‡ Swiss Federal Institute of Technology Zurich. 10.1021/es702959a CCC: $40.75
Published on Web 04/29/2008
2008 American Chemical Society
Fontela et al. (7), Castiglioni et al. (8), and Hummel et al. (9) detected illicit psychoactive drugssincluding 3,4-methylenedioxymethamphetamine (MDMA or ecstasy), 3,4-methylenedioxyethamphetamine (MDEA), morphine, methadone, as well as their metabolitessin treated wastewaters and the receiving aquatic environments. Municipal WWTPs were identified as the main source of human pharmaceuticals in the environment (1, 2, 10–14). The fate of psychoactive drugs in the aquatic environment, particularly in river and stream sediments, remains poorly understood. This is at least in part due to the lack of analytical methods for quantification of these drugs in sediment samples. Concentration data in sediments needs to compliment data on aqueous phase concentrations, particularly in light of the high susceptibility of psychoactive drugs to microbial (and abiotic) degradation. However, analytical methods for other pharmaceuticals such as antibiotics (i.e., fluoroquinolones, sulfonamides, tetracyclines), estrogens, and nonsteroidal anti-inflammatory drugs (15) (16) (17) (18) in sediments have already been developed, and there are several published studies on the fate of these compounds in soils and sediments (19–25) and during groundwater recharge and riverbank filtration (26, 27). With regard to the fate of psychoactive drugs in aquatic environments, no information is yet available on the sorption of these pollutants to sediments. Extensive H-bond interactions between the sorbents and these drugs are likely given the high polarity of these chemicals. Many of these drugs also contain functional groups that undergo acid-base chemistry. Consequently, electrostatic interactions between sorbate and sorbent, most likely between cationic species of the drug and negatively charged sorption sites in/on the sorbent, may play an important role in overall sorption for psychoactive drugs that are predominantly cationic at neutral pH. The latter include chemicals from the opiates and opioides classes. Electrostatics were shown to explain strong sorption of some polar (i.e., KOW < 1) pharmaceuticals, including ciprofloxacin and enrofloxacin to biosolids and soils (28), and of sulfonamide antibiotics, to natural organic matter and sludge (29, 30). Similarly, the cationic species of sulfonamides show higher affinities than the neutral species for negatively charged clay minerals and clay mineral-humic complexes (31). The impact of psychoactive pharmaceuticals on the aquatic ecosystems is unclear (32–34). For most of the psychoactive drugs, the ecotoxicological database is too weak to perform a comprehensive environmental risk assessment (ERA). A few studies indicate severe impacts on benthic organisms upon exposure to the antiepileptic carbamazepine (35). Advancing toward a proper assessment of potential impacts on sediment-dwelling organisms requires (i) an analytical method to quantify these drugs in sediments; (ii) data on the sorption affinities of selected psychoactive drugs, ideally from different classes of psychoactive drugs, to river and stream sediments; and (iii) information on the kinetics and thermodynamics of the desorption of these drugs from sediments. The latter information is important as desorption from sediments increases the bioavailability of the drugs. There are two major objectives to this work. The first is to develop an analytical method to quantify important psychoactive drugs from several chemical classes—including antiepileptics, opiates, opioids, and tranquilizers, as well as two principle metabolites—in river and stream sediments. The second objective is to collect an extensive data set on (i) the sorption kinetics of these compounds to sediments versus their degradation kinetics and (ii) the concentration VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6415
TABLE 1. Chemical Structures of the Investigated Pharmaceuticals
dependence of sorption and its reversibility. To this end, batch experiments were performed to determine the sorption kinetics and the apparent sorption-desorption equilibria of nine chemically diverse psychoactive drugs (some are also metabolites) and three metabolites to two nonsterilized natural river sediments that differed in their physicochemical characteristics (i.e., organic matter content and particle size fractionation).
Materials and Methods Chemicals and Standards. Table 1 provides chemical structures and CAS registry numbers of the psychoactive drugs and the sulfonamide antibiotics investigated in the current study. The standards were obtained from different sources as described in the Supporting Information (Chapter 3). Table 2 shows which surrogate standard was used for each target compound. Analysis of Sediment Samples. We developed an analytical method for the analysis of psychoactive drugs in sediments based on the method for aqueous samples by Hummel et al. (9). In brief, sediment samples were extracted by pressurized 6416
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
solvent extraction (PLE) (ASE-200, Dionex, Idstein, Germany). Sediment (1 g) was filled into an 11 mL stainless steel extraction cell from Dionex (Idstein, Germany). The sediment was thoroughly mixed with about 10 g of quartz sand, followed by spikes of 100 ng of each surrogate standard to the mixture. The extraction solvent was methanol/water (50/50; v/v). The operating conditions were: extraction temperature, 100 °C; extraction pressure, 100 bar; preheating period, 5 min; static extraction period, 5 min; number of extraction cycles, 3; solvent flush, 120% of cell volume; nitrogen purge, 30 s. The final extract (approximately 35 mL) was transferred (rinsed with about 80 mL of groundwater-water in 2-3 portions) to a 1000 mL volumetric flask, followed by addition of 500 mL of groundwater. The groundwater was collected from a well in Arenberg near Koblenz, Germany. Measurements showed that the groundwater was pristine and free of all targeted analytes. The quality of the groundwater is described in Table S2 of the Supporting Information. Samples were subsequently cleaned, and analytes were preconcentrated by solid phase extraction (SPE), followed LC tandem MS detection for separation and quantification. More details are
VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6417
1 2 1
2 2 1 1
1 2
Tranquilizers diazepame oxazepamf temazepamf
Opiates, opioids morphineg codeineh dihydrocodeinei tramadola
Antibiotics sulfamethoxazolej N4-acetylsulfamethoxazolek 104 ( 4 96 ( 6
115 ( 14b 74 ( 3b 89 ( 4b 74 ( 18b 2 5
5 5 5 5
2 5 2
5
2 2
LOQ [ng/g]
32 ( 6 57 ( 4
76 ( 1 82 ( 8 81 ( 7 51 ( 7
43 ( 11 74 ( 7 76 ( 5
71 ( 7
61 ( 5 61 ( 6
absolute recovery [%]
133 ( 3 114 ( 5
98 ( 19 104 ( 16 103 ( 15 96 ( 10
112 ( 8 109 ( 2 111 ( 6
134 ( 8
114 ( 4 115 ( 8
relative recovery [%]
c
Diazepam d5, f Oxazepam d6,
c e
c c
c
d
d
d
d
101 105 50
101 87 128
86
93 85
Emscher
d
96 129 105
d
101 93
Landgraben
c
102 98 102 133
97 77 87
137
129 103
Schwarzbach
relative recovery [%]
g
h
c
c
Morphine d6,
112 97
c
c
c
Codeine
114 102
c
c
c
c
c
c
c
c
c
c
c
c
c c
Murn c
Schlauengraben
various sediments (N ) 1) spiked concentration: 50 ng
a Surrogate standards: carbamazepine 13C15N. b From ref. 9. c Not spiked. d Evaluation not possible. Surrogate standards: d6, i Benzoylecgonine d8, j Sulfamethoxazole d4, k N4-Acetylsulfamethoxazole d6.
52 ( 15 138 ( 21
86 ( 8b 76 ( 5b 92 ( 5b 74 ( 9b
95 ( 14b 102 ( 7b 78 ( 17b
124 ( 12b
94 ( 13b
2 84 ( 5b 93 ( 9b 78 ( 4b
101 ( 7b 87 ( 3b
relative recovery [%]
76 ( 7b 66 ( 5b
absolute recovery [%]
1 1
LOQ [ng/mL]
Antiepileptics carbamazepinea dihydro-carbamazepinea dihydroxy-dihydrocarbamazepinea
substances
aqueous phase (N ) 4)
reference sediment (N ) 3) spiked concentration: 100 ng
TABLE 2. LOQ, Absolute and Relative Recoveries ± Confidence Intervals (P = 95%) of the Analytes in Aqueous Phase and Different Sediment Matrices
provided in the Supported Information and in Hummel et al. (9). Retention times and masses can be found in Supported Information Table S3. In addition to the psychoactive drugs, sulfamethoxazole and its metabolite N4-acetylsulfamethoxazole were included in the measurements and were detected according to Go¨bel et al. (36). Analysis of Water Samples. The labeled surrogate standards (100 ng/mL) were spiked to 1 mL of particle-free water withdrawn from centrifuged batch reactors. The water (50 µL) was directly injected into the LC electrospray tandem MS. The psychoactive drugs and sulfonamides were detected in positive mode. Details are given by Hummel et al. (9). The sulfonamides were analyzed as described in the previous section. Method Validation. Recoveries and Limit of Quantification. Individual recoveries for each analyte over the entire method were determined by spiking natural sediments prior to PLE with analytes and surrogate standards. Absolute compound-specific recoveries were determined by comparing the peak area of a sediment extract (MeOH/MilliQ water, 1:9, v/v), which went through the whole method, to the area obtained from detecting a standard solution prepared in the same solvent mixture as the sample extract. Relative recoveries were calculated by correcting absolute recoveries with the recoveries of the spiked surrogate standards. The limit of quantification (LOQ) is referred to the second lowest calibration point in the linear regression as long as the respective calculated signal/noise ratio of the analyte in the native samples extracts was at least 10. Blank samples (without analytes) and control samples (without sediment, but spiked with analytes) were included in each samples series. Matrix Effects. Ion suppression or enhancement during electrospray ionization due to matrix effects is a major cause for incorrect quantitative results. To assess the influence of matrix components, aliquots of final sample extracts were spiked with the analytes prior to analysis to determine recoveries excluding any losses caused by sample preparation such as SPE. Ion suppression was calculated by comparing the difference in analyte quantities determined from the spiked and nonspiked final sediment extract to the spiked quantity. Batch Experiments. Experiments were carried out with two sediments that differed in total organic carbon contents (TOC) and grain size distributions. The Burgen sediment (TOC: 0.74%, clay/silt: 10%) and the Dausenau sediment (TOC: 4.36%, clay/silt: 47%) were collected from the stream Baybach (a tributary of the Mosel River) near Burgen and the stream Unterbach (a tributary of the Lahn River) near Dausenau, respectively. Exact sampling locations and physicochemical characteristics of the sediments are provided in Figure S1 and Table S1 of the Supporting Information, respectively. Both sampling sites are characterized by low anthropogenic impacts. Sediment concentrations of target analytes were below the LOQ. Air-dried sediments with particle sizes
