Aquatic integrative passive samplers are used to determine aqueous concentrations of polar organic pollutants, yet their uptake mechanisms are poorly understood. We introduce a one-dimensional model ...to simulate uptake by a passive sampler, Chemcatcher. The model considers the uptake as molecular diffusion through a series consisting of the aqueous boundary layer (ABL), the membrane filter (MF), and the sorbent disk with concurrent sorption by matrix of the MF and the disk. Uptake profiles of ∼20 polar chemicals measured over a week and a month were accurately modeled. Characteristic behaviors such as lag phases, linear and curved uptake, and equilibrating behavior were explained well by the model. As the model is mechanistically based, it was able to show the combined influences of the MF/water (K MF/w) and disk/water (K disk/w) partition coefficients, diffusion coefficients, and the ABL thickness on the sampling rates. On the basis of the model results, we offer three concrete recommendations for achieving the linear uptake needed for measuring time-weighted average concentrations: (i) use a MF that does not significantly sorb chemicals (e.g., log K MF/w < 3) to avoid lag phases, (ii) use a sorbent with strong sorption properties (e.g., log K disk/w > 6) for effective trapping of chemicals on the disk top layer, and (iii) make the ABL and/or the MF thicker so that the diffusion toward the disk slows.
Food contact materials (FCM) may contain complex mixtures of estrogenic chemicals. A yeast estrogen screen performed on high performance thin-layer chromatography plates (planar-YES, P-YES) is ...promising for analysis of such mixtures, as it could allow for better elucidation of effects compared with established methods in microtiter plates. However, the P-YES has not been directly compared with established methods. We compared the performance of a microtiter plate YES (lyticase-YES, L-YES) to P-YES on silica gel HPTLC plates using 17β-estradiol (E2), 20 chemicals representative of migrants from plastic FCM, and three migrates of coated metal food cans. Effective doses (ED
10
, ED
50
) and estradiol equivalencies were calculated for each chemical. Thirteen chemicals had calculable EDs in the L-YES or P-YES, with average EDs 13-fold (range 0.63–36) more potent in P-YES than in the L-YES. Normalized to E2, the median estrogenicity was within 1.5-fold (0.43–8.8) between the assays. Therefore, P-YES was as or more sensitive than L-YES but potencies relative to E2 were comparable between assays. With chromatography, the P-YES detected estrogenicity in coated metal cans, effects that were unmeasurable in L-YES. With the sample preparation methods used in this study, both YES assays are sufficiently sensitive to detect bisphenol A below the specific migration limit for plastic packaging (0.05 mg/kg food). This study demonstrates that P-YES outperforms L-YES because it is more sensitive, provides comparable estradiol equivalents, and circumvents confounding mixture effects. The P-YES will be useful for routine monitoring of FCM and toxicant identification in problematic materials.
Graphical abstract
Ecotoxicological screening of surface waters can involve multiple analyses using multiple bioassay and chemical analytical methods that require enriched samples to reach low concentrations. Such ...broad screening of the same sample necessitates sufficient sample volume—typically several liters—to produce a sufficient amount of enriched sample. Often, this is achieved by performing parallel solid-phase extractions (SPE) where extracts are combined into a pool—this is a laborious process. In this study, we first validated our existing SPE method for the chemical recovery of an extended set of compounds. We spiked four estrogenic compounds and 11 herbicides to samples from independent rivers (1 L) and wastewater treatment plant effluents (0.5 L). Then, we investigated the effect of increased sample loading of the SPE cartridges on both chemical and biological recoveries by comparing the validated volumes with four times larger sample volumes (i.e., 4 L river water and 2 L effluent). Samples were analyzed by LC-MS/MS and three bioassays: an estrogen receptor transactivation assay (ERα-CALUX), the combined algae test, and a bacterial bioluminescence inhibition assay. Our existing SPE method was found to be suitable for enriching the extended set of estrogens and herbicides in river water and effluents with near to perfect chemical recoveries (~ 100%), except for the herbicide metribuzin (46 ± 19%). In the large volume river and effluent samples, the biological activities and concentrations of the spiked compounds were between 87 and 104% of those measured with the lower sample loading, which is adequate. In addition, the ratio between the large and original volume SPE method for the non-target endpoint (bacterial bioluminescence inhibition) was acceptable (on average 82 ± 9%). Results indicate that our current water extraction method can be applied to up to four times larger sample volumes, resulting in four times more extract volumes, without significant reductions in recoveries for the tested estrogens and herbicides.
Graphical abstract
ᅟ
Passive samplers for polar organic compounds often use a polyethersulfone (PES) membrane to retain the particulate sorbent material (e.g., in a POCIS; polar organic chemical integrative sampler) or ...to reduce the sampling rate and thus extend the kinetic regime (e.g., in a Chemcatcher). The transport kinetics over the PES membrane are evaluated here in a short-term (6 days) and a long-term (32 days) experiment with POCIS and Chemcatchers. Passive samplers were placed in a channel with flowing river water that was spiked with 22 organic chemicals including pharmaceuticals, pesticides and biocides; with logK ow (logarithmic octanol–water partitioning coefficient) values between −2.6 and 3.8. Samplers were removed at intervals and membranes and sorbent material were extracted and analyzed with LC-MS/MS. Uptake kinetics of the compounds fell between two extremes: (1) charged chemicals and chemicals of low hydrophobicity did not accumulate in PES and rapidly transferred to the sorbent (e.g., diclofenac) and (2) more hydrophobic chemicals accumulated strongly in the PES and appeared in the sorbent after a lag-phase (e.g., diazinon and diuron). Sorption kinetics were modeled with a three-compartment first-order kinetic model to determine uptake and elimination rate constants and partitioning coefficients. Water PES partitioning coefficients fitted with the model correlated well with experimentally determined values and logK ow. Sampling rates of Chemcatcher (0.02–0.10 L/d) and POCIS (0.02–0.30 L/d) showed similar patterns and correlated well. Thus the samplers are interchangeable in practical applications. Longer lag-phases may pose problems when calculating time-weighted average aqueous concentrations for short passive sampling windows and for a correct integrative sampling of fluctuating concentrations.
Display omitted
•Measured and modelled concentration profiles systematically evaluated.•Variable concentrations do not explain observed deviations from true concentrations.•Longer passive sampling ...periods or early/late peak lead to higher uncertainty.•Event-based passive sampling in sewers comparable to sampling <10 min.
Considerable pollutant loads can enter surface waters during rain events. Three factors challenge quantification of these pollutant fluxes using traditional sampling methods: (i) concentration fluctuations; (ii) unknown event duration; and (iii) placement, operation, and maintenance of equipment. Passive samplers offer the advantage of sampling in a continuous mode without power supply. However, variable uptake rates due to environmental factors and desorption in the case of fluctuating concentrations can affect the accuracy of time-weighted average (TWA) concentration estimates. While uncertainties related to environmental factors could be accounted for with additional effort, we can neither control nor quantify the concentration variability. We present measured and modelled concentration profiles at high temporal resolution and provide a systematic approach to assessing deviations from true TWA concentration due to fluctuating concentration profiles. We evaluate sampling of sewer overflows (0.3–14 h) with Chemcatcher and 1-week sampling in rivers. The uncertainty due to fluctuating concentrations is small, and other factors such as chemical analyses and sampler calibration have a similar or higher impact. The uncertainty due to fluctuations clearly increases with the sampling duration, particularly when exceeding the half-life of equilibrium. We conclude that passive sampling can also be used in wastewater systems with potentially high concentration variations.
Wastewater treatment plant effluents are important point sources of micropollutants. To assess how the discharge of treated wastewater affects the ecotoxicity of small to medium-sized streams we ...collected water samples up- and downstream of 24 wastewater treatment plants across the Swiss Plateau and the Jura regions of Switzerland. We investigated estrogenicity, inhibition of algal photosynthetic activity (photosystem II, PSII) and growth, and acetylcholinesterase (AChE) inhibition. At four sites, we measured feeding activity of amphipods (Gammarus fossarum) in situ as well as water flea (Ceriodaphnia dubia) reproduction in water samples. Ecotoxicological endpoints were compared with results from analyses of general water quality parameters as well as a target screening of a wide range of organic micropollutants with a focus on pesticides and pharmaceuticals using liquid chromatography high-resolution tandem mass spectrometry. Measured ecotoxicological effects in stream water varied substantially among sites: 17β-estradiol equivalent concentrations (EEQbio, indicating the degree of estrogenicity) were relatively low and ranged from 0.04 to 0.85 ng/L, never exceeding a proposed effect-based trigger (EBT) value of 0.88 ng/L. Diuron equivalent (DEQbio) concentrations (indicating the degree of photosystem II inhibition in algae) ranged from 2.4 to 1576 ng/L and exceeded the EBT value (70 ng/L) in one third of the rivers studied, sometimes even upstream of the WWTP. Parathion equivalent (PtEQbio) concentrations (indicating the degree of AChE inhibition) reached relatively high values (37 to 1278 ng/L) mostly exceeding the corresponding EBT (196 ng/L PtEQbio). Decreased feeding activity by amphipods or decreased water flea reproduction downstream compared to the upstream site was observed at one of four investigated sites only. Results of the combined algae assay (PSII inhibition) correlated best with results of chemical analysis for PSII inhibiting herbicides. Estrogenicity was partly and AChE inhibition strongly underestimated based on measured steroidal estrogens respectively organophosphate and carbamate insecticides. An impact of dissolved organic carbon on results of the AChE inhibition assay was obvious. For this assay more work is required to further explore the missing correlation of bioassay data with chemical analytical data. Overall, the discharge of WWTP effluent led to increased estrogenicity, PSII and AChE inhibition downstream, irrespective of upstream land use.
Passive sampling is a well-established tool for monitoring time-weighted average concentrations of polar and semi-polar organic contaminants in streams at flow velocities between 0.1 and 0.4 m s−1. ...However, its application under low-flow conditions (10−5 to 0.01 m s−1) – as encountered in hyporheic zones – has been scarcely reported. In this study, 3 novel passive sampler configurations were developed for the monitoring of (semi-)polar organic pollutants and related transformation products across the water-sediment interface and thus across varying hydrodynamic conditions. Their design was inspired by Chemcatcher and diffusive gradients in thin films for organics. To determine the most optimal sampler design, an uptake experiment was completed involving the 3 novel passive sampler configurations and a reference Chemcatcher in polar configuration. The experiments consisted of a circular flume that simulated the main channel of a stream and an aquarium with stagnant water that represented the underlying hyporheic zone. The systems were exposed to 192 organic pollutants at environmental concentrations, and the samplers were then collected, extracted and analyzed using liquid chromatography high-resolution mass spectrometry after 2, 6 and 14 days. The configuration that was most insensitive to different hydrodynamic conditions consisted of a reversed-phase sulfonated styrenedivinylbenzene disk as the receiving phase that was covered by an agarose diffusion gel and topped with a polyethersulfone membrane filter. To further evaluate its environmental application, samplers were installed downstream of a sewage treatment plant located at an urban stream in Berlin, Germany (Erpe). The samplers were mounted on custom-made holders which were subsequently embedded in the stream bed to position samplers above (0.30 m) and within the sediment (−0.15/-0.30/-0.45 m) for 11 days. Target and suspect screening workflows were then applied to identify common concentration patterns and link parent attenuation to transformation product formation. A total of 104 concentration profiles were determined, suggesting the efficiency of the proposed sampling strategy in the water-sediment interface. Valsartan acid was the only known transformation product indicative of hyporheic zone-driven attenuation as its concentration in porewater by far exceeded its concentration in surface water. Similar patterns were observed for a larger list of suspected transformation products, of which a sotalol transformation product was tentatively identified. Overall, the established sampling methodology can be effectively used to quantify organic contaminants during low-flow conditions and is suitable for the characterization of attenuation patterns of organic pollutants in hyporheic zones.
Display omitted
•Design of 3 novel passive sampler configurations for low-flow conditions.•Broad collection of low-flow passive sampling rates for 144 organic contaminants.•Passive sampling of organics across the water-sediment interface of an urban stream.•Attenuation patterns of 104 organic contaminants in the hyporheic zone.•Detection of transformation products as indicators for biological degradation.
In a large field study, the in-situ calibration of the Chemcatcher® passive sampler – styrenedivinylbenzene (SDB) covered by a polyether sulfone (PES) membrane – was evaluated for 322 polar organic ...micropollutants. Five rivers with different agricultural and urban influences were monitored from March to July 2012 with two methods i) two-week time-proportional composite water samples and ii) two-week passive sampler deployment. All substances – from different substance classes with logKow −3 to 5, and neutral, anionic, cationic, and zwitterionic species – were analyzed by liquid-chromatography high-resolution tandem mass spectrometry. This study showed that SDB passive samplers are well-suited for the qualitative screening of polar micropollutants because the number of detected substances was similar (204 for SDB samples vs. 207 for composite water samples), limits of quantification were comparable (median: 1.3 ng/L vs. 1.6 ng/L), and the handling in the field and laboratory is fast and easy. The determination of in-situ calibrated sampling rates (field Rs) was possible for 88 compounds where the R2 from the regression (water concentration vs. sampled mass on SDB disk) was >0.75. Substances with moderately fluctuating river concentrations such as pharmaceuticals showed much better correlations than substances with highly fluctuating concentrations such as pesticides (R2 > 0.75 for 93% and 60% of the investigated substances, respectively). Flow velocity (0.05–0.8 m/s) and temperature (5–20 °C) did not have an evident effect on the field Rs. It was observed that ionic species had significantly lower field Rs than neutral species. Due to the complexity of the different transport processes, a correlation between determined field Rs and logDow could only predict Rs with large uncertainties. We conclude that only substances with relatively constant river concentrations can be quantified accurately in the field by passive sampling if substance-specific Rs are determined. For that purpose, the proposed in-situ calibration is a very robust method and the substance specific Rs can be used in future monitoring studies in rivers with similar environmental conditions (i.e., flow velocity, temperature, pH).
Display omitted
•SDB passive sampler perfectly suitable for qualitative screening of micropollutants.•Broad substance spectrum detected, detection limits comparable with water samples.•Robust field sampling rates were determined for 88 substances by in-situ calibration.•Substances with highly fluctuating water concentrations are more difficult to quantify.•Effects of flow velocity and temperature on field sampling rates were not evident.