Equilibrium Partition Coefficients of Diverse Polar ... - ACS Publications

ARTICLE pubs.acs.org/est

Equilibrium Partition Coefficients of Diverse Polar and Nonpolar Organic Compounds to Polyoxymethylene (POM) Passive Sampling Devices Satoshi Endo,*,† Sarah E. Hale,‡ Kai-Uwe Goss,†,§ and Hans Peter H. Arp‡ †

Department of Analytical Environmental Chemistry, UFZ
Helmholtz Centre for Environmental Research, Permoserstrasse 15, D-04318 Leipzig, Germany  ‡ Department of Environmental Engineering, Norwegian Geotechnical Institute (NGI), P.O. Box 3930, Ulleval Stadion, N-0806, Oslo, Norway § Institute of Chemistry, University of Halle-Wittenberg, Kurt-Mothes-Strasse 2, D-06120 Halle, Germany

bS Supporting Information ABSTRACT: Equilibrium passive samplers (EPS) based on polyoxymethylene (POM) are increasingly used for determining freely dissolved water and pore water concentrations of hydrophobic organic compounds in the environment. Unlike other polymeric materials commonly used as EPS, namely poly(dimethylsiloxane) (PDMS) and low-density polyethylene (PE), POM is a polar polymer, containing repeating H-bond accepting ether units. Thus, POM is expected to be a more sensitive EPS than PDMS and PE for polar, H-bond donating compounds, such as many hormones, pharmaceuticals, and biocides. To better characterize the sorption capacity of POM for diverse polar and apolar compounds, equilibrium POM
water partition coefficients, KPOM/w, were measured for 56 compounds, including several classes of polar compounds and organochlorine pesticides. Using this data set and literature data, various POM-partitioning models were calibrated and validated for their ability to predict KPOM/w. The best performing models tested were an Abraham descriptor based polyparameter linear free energy relationship (PP-LFER) (SD = 0.24 log units) and COSMOthermX (SD = 0.37 log units). The performance of SPARC (SD = 0.61 log units) and log
log correlations with Kow (SD = 0.49 log units) were lower. A comparison with PDMS and PE confirmed expectations that POM exhibits a higher sensitivity for H-bond donating polar compounds than PDMS and PE do for these compounds. These findings expand the domain of chemicals for which POM can be used as an EPS sampler, and demonstrate that POM is as suitable a passive sampler for many polar organic compounds as it is for hydrophobic organic compounds.

’ INTRODUCTION Equilibrium passive samplers are increasingly being utilized as a tool to determine freely dissolved water and pore water concentrations of hydrophobic organic compounds as well as their diffusion gradients across various environmental compartments.1
7 In contrast to the conventional method for measuring water phase concentrations (i.e., sampling of water followed by filtration and extraction), equilibrium passive sampling methods can avoid laborious water sampling procedures and potential artifacts due to the presence of dissolved organic matter. The principle of equilibrium passive sampling is that a device comprised of a polymer is deployed in an environmental phase over a certain period of time that is sufficient for sorption equilibrium to be established between the sampler polymer and the surrounding phase. The device is retrieved, the concentration of the target chemical in the polymer phase (Cpolymer) is measured, and the freely dissolved aqueous phase concentration (Cw) is calculated r 2011 American Chemical Society

using the following relationship: Cw ¼ Cpolymer =Kpolymer=w

ð1Þ

where Kpolymer/w is the partition coefficient of the chemical between the polymer and water. Thus, the chemical-specific Kpolymer/w has to be known to infer Cw, and the accuracy of the used value of Kpolymer/w is directly related to the accuracy of the Cw measurements. From eq 1 the higher the value of Kpolymer/w, the higher the concentration in the polymer after sampling, and thus the higher the sensitivity for Cw measurements. Models that can predict Kpolymer/w for a passive sampler would be highly useful. Currently, few prediction models exist for passive Received: August 18, 2011 Accepted: October 17, 2011 Revised: October 13, 2011 Published: October 17, 2011 10124

dx.doi.org/10.1021/es202894k | Environ. Sci. Technol. 2011, 45, 10124–10132

Environmental Science & Technology samplers, and Kpolymer/w has to be measured through timeconsuming sorption experiments for chemicals of interest. A high performing Kpolymer/w predictive model would allow for assessing whether the Kpolymer/w is high enough to achieve the desired sensitivity for a given compound, estimating how much polymer material should be used when deployed in soils and sediments (to minimize depletion artifacts), and potentially would allow for skipping the necessity of preliminary experiments for Kpolymer/w determination (presuming the model is accurate enough) and thereby enabling Cw determinations for a wider range of compounds than otherwise possible. Moreover, estimations are also useful to assess the quality of experimentally determined Kpolymer/w. Polyoxymethylene (POM)-based equilibrium passive sampler devices1 are commonly deployed in sediment
pore water systems2,3 in both marine and freshwater systems.4,5 POM has an excellent physical and chemical stability and a smooth and hard surface that renders it less susceptible to the trapping of particles or biofouling compared to other samplers with rough or sticky surfaces.1 As an equilibrium passive sampler, POM exhibits sufficiently large partition coefficients for hydrophobic compounds such as polychlorinated biphenyls (PCBs),2,4 polycyclic aromatic hydrocarbons (PAHs),1,4,6 and polychlorinated dibenzodioxins/furans,7 allowing for the determination of Cw at pg/L levels. In contrast to the other most frequently used equilibrium passive sampler sorbents, poly(dimethylsiloxane) (PDMS) and low-density polyethylene (PE), POM contains a repeating polar group (
CH2
O
CH2
) in its molecular structure. Thus, it is anticipated that POM has a higher affinity for some polar chemicals than PDMS and PE. So far, however, POM has not been used to extract contaminants with polar functional groups. Accordingly, POM
water partition coefficients (KPOM/w [Lwater/kgPOM]) have not been reported for any polar chemicals. The purposes of this study are (i) to determine KPOM/w for diverse neutral polar and apolar organic compounds, (ii) to calibrate and evaluate estimation models for KPOM/w, and (iii) using the calibrated estimation models obtained, to compare POM with other polymers used for equilibrium passive sampling with regard to their affinity for a diverse selection of organic compounds.

’ MATERIALS AND METHODS POM Sampler. A thin sheet of POM (76 μm thick) purchased from CS Hyde Company (Lake Villa, IL) was used, as this commonly available POM product has been frequently used as an equilibrium passive sampler.2,6 POM, as received, was cut into strips and immersed in isohexane or heptane for at least 1 day at room temperature and then in methanol for at least another day to remove impurities. Data Sets of KPOM/w. Two experimental data sets of KPOM/w are presented in this study. The first set was assembled with the intention to use it for calibrating estimation models (thus referred to as the “calibration set”). Chemicals for the calibration set were selected based on the chemical diversity and the availability of compound descriptors (see the explanation below for “polyparameter linear free energy relationships”) and were measured at the UFZ
Helmholtz Centre for Environmental Research in Leipzig. The methods for this experiment are described in the next section Batch Experiments for Calibration Compounds. In addition to our own

ARTICLE

measured data, experimental KPOM/w values of 53 PCBs2 and of 20 PAHs6 from other studies were included in the calibration data set (data listed in Table S1 in the Supporting Information). In these cited studies, the KPOM/w values were measured using the same POM product as used here. In the literature there exist more PAH, PCB and dioxin data for KPOM/w determined using differently prepared POM samplers (e.g., refs 1, 4, 7
9). These data were not considered here, as particularly for the calibration data set we sought to control for any possible influences of sampler preparation. The second set, referred to as the “validation set”, consists of those compounds that are of environmental concern but whose available compound descriptors (see below for the descriptors used) are less certain than for the calibration compounds. These validation compounds include organochlorine pesticides (OCPs) and were measured at the Norwegian Geotechnical Institute (NGI) in Oslo. The methods for this experiment are described in the section Batch Experiments for Pesticide Validation Compounds. Moreover, KPOM/w values for 10 (isomer unspecified) alkylated PAHs from ref 6 were also included in the validation data set. Batch Experiments for Calibration Compounds. Chemicals were obtained from different providers. The purity was >95%. Experimental water was treated with a Milli-Q A10 Ultrapure Water Purification System (Millipore, Billerica, MA). NaN3 (200 mg L
1) was dissolved in water to minimize microbial activities during the batch experiments. For the experiments with ionizable compounds such as phenols and anilines, a 10 mM phosphate buffer solution with the pH value adjusted to 7.40 was used. All KPOM/w values measured represent the values for the neutral species. A cleaned POM strip (10 or 20 mg) was weighed into a 20-mL crimp-top vial, which then received 10
20 mL of water containing the electrolytes mentioned above. The test batch was spiked with a methanol stock solution containing 3
6 calibration compounds. The methanol content in water was e0.1% (v/v). Five replicates were prepared. Only one concentration for each chemical was examined, as linear sorption to POM has been reported from the pg/L range and upward.1,2,6 The vials were immediately sealed with an aluminum/silicon- or PTFE/silicon-lined cap and were shaken horizontally for 28 d at room temperature (25 ( 2 °C). This length of time has been shown to be sufficient for equilibration of PCBs2 and PAHs,6 which—due to their high sorption coefficients and low diffusivities in condensed phases—likely exhibit a longer equilibrium time than the compounds measured here. A model calculation also shows that 28 d should be sufficient.10 For volatile compounds, the headspace of the vial was analyzed with GC/MS and compared with calibration curves from that of freshly prepared standard solutions to quantify the aqueous phase concentration. The instrumental conditions for the headspace-GC/MS measurements are described in the Supporting Information. Following headspace sampling, the cap of the vial was removed and the POM strip was retrieved. The POM was dried with clean tissue and extracted with 1 mL of isohexane or ethyl acetate. The extract was analyzed with GC/MS to determine the concentration in POM (for details about the GC measurements on extracts, see ref 11). For nonvolatile compounds, the headspace measurement was not conducted, but, instead, an aliquot of the aqueous phase was liquid
liquid extracted after equilibration and analyzed to determine the aqueous phase concentration following the method described in ref 11. Because the concentrations in both the POM and water phases were measured, KPOM/w can be obtained even if there was a mass loss of the analyte during the experiment (e.g., volatilization, 10125

dx.doi.org/10.1021/es202894k |Environ. Sci. Technol. 2011, 45, 10124–10132

Environmental Science & Technology

ARTICLE

transformation, glassware sorption). Nevertheless, if the total recovery (POM + water + headspace) was KPDMS/w and KPOM/w > KPE/w, agreeing with the fact that POM has H-bond acceptor sites in its molecular structure and PDMS and PE do not. These H-bond donor compounds include substituted phenols, bisphenol A, natural and synthetic estrogens, and various pharmaceuticals. Note that the solute H-bond acceptor property (B) does not appear to influence the relative sorption coefficients to these sorbents, as none of POM, PDMS, nor PE contain an H-bond donor group. In Figure S8, a similar comparison is presented between KPOM/w and partition coefficients to polyacrylate (KPA/w). Interestingly, KPOM/w values are similar to KPA/w and there was no systematic trend for particular types of compounds. These sorbents appear to share similar interaction properties, as is also indicated by the

similarity of their PP-LFER parameters (Table 3). Moreover, KPOM/w is consistently lower than KPA/w (by a factor of 4 on average). This can be because there are crystalline regions in POM that are not available for partitioning, whereas PA used for solid phase microextraction seems much more amorphous and largely available for partitioning. Implications for Passive Sampling. As confirmed in this and previous work,11,28 PP-LFER models are well suited for fitting experimental values of Kpolymer/w for a number of compounds with varying size and polarity. The accuracy is typically a SD of 0.2 log units. If high-quality descriptors are available, the use of PPLFERs is so far the most reliable method to predict Kpolymer/w. COSMOthermX also calculates partition coefficients with fairly high accuracy, typically within 1 log unit from the measured values. Because COSMOthermX does not need any compoundspecific descriptor, it is an attractive tool for the first estimation of partition coefficients. Thus, these two models are recommended for estimating log KPOM/w, though the consistent outlying compound classes identified here should be kept in mind as, if not accountable to experimental artifacts, they may indicate the limit of the chemical application domain of these models. For the SP-LFER log Kow correlation and the SPARC model, worse correlations resulted and thus the size of the chemical application domain is more limited, particularly if high accuracy is required. We emphasize that when using either experimental or estimated log KPOM values, the reported error should be accounted for. KPOM/w, KPDMS/w, and KPE/w for hydrophobic aromatic compounds (e.g., PAHs) fall within a narrow range. Thus, POM, PDMS, and PE are equivalent in terms of the affinity for these chemicals. POM is advantageous if one aims to extract H-bond donor polar compounds from environmental phases. For example, Sacks et al.33 used PE to passively sample alkylphenols and triclosan. These chemicals possess a
OH group in the structure and thus H-bond donor properties. Hence, POM is expected to be a more sensitive sorbent for these compounds, as it would exhibit high log Kpolymer/w and therefore more sensitive measurements. One aspect to consider here is that a particular type of POM product was used in this work. An area of research not considered here or in depth in other publications is the potential impact on partition coefficients from temperature, material fabrication (e.g., presence of plasticizer, molding vs weaving), and, in particular, crystallinity.22,23 This research is recommended for future work, and is ongoing in our laboratories. Moreover, the time necessary 10130

dx.doi.org/10.1021/es202894k |Environ. Sci. Technol. 2011, 45, 10124–10132

Environmental Science & Technology for equilibration could also depend on the types and conditions of POM products. POM samplers in general need relatively long equilibration times due to slow kinetics. For example, an equilibration time of at least 28 d was recommended for PCBs and PAHs for the POM material used here.2,6 If a POM sampler is deployed for a time period that is too short, it may not reach equilibrium and could register an erroneous concentration. Thus, great attention should be paid to sufficient equilibration time. Also, because of long equilibration times with the material used here, and lack of knowledge about uptake kinetics under different flow regimes, POM is currently not suitable as a passive sampler to monitor highly fluctuating concentrations (e.g., in wastewater); how passive samplers could be developed and optimized for such environments is an interesting topic for research. From a practical point of view, it is recommended to use a POM material that has been well characterized to avoid these potential influences. The commonly available POM material tested here shows good potential for applications as an equilibrium passive sampler for H-bond donating compounds, including many hormones, additives, biocides, and pharmaceuticals.

’ ASSOCIATED CONTENT

bS

Supporting Information. Experimental and calculated partition coefficients, PP-LFER descriptors used, and additional figures and tables. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: +49 341 235 1818; fax: +49 341 235 1443; e-mail: [email protected].

’ ACKNOWLEDGMENT We acknowledge valuable discussions with Steven B. Hawthorne (Energy and Environmental Research Center, University of North Dakota). ’ REFERENCES (1) Jonker, M. T. O.; Koelmans, A. A. Polyoxymethylene solid phase extraction as a partitioning method for hydrophobic organic chemicals in sediment and soot. Environ. Sci. Technol. 2001, 35 (18), 3742–3748. (2) Hawthorne, S. B.; Miller, D. J.; Grabanski, C. B. Measuring low picogram per liter concentrations of freely dissolved polychlorinated biphenyls in sediment pore water using passive sampling with polyoxymethylene. Anal. Chem. 2009, 81 (22), 9472–9480. (3) Cornelissen, G.; Pettersen, A.; Broman, D.; Mayer, P.; Gijs, D. B. Field testing of equilibrium passive samplers to determine freely dissolved native polycyclic aromatic hydrocarbon concentrations. Environ. Toxicol. Chem. 2008, 27 (3), 499–508. (4) Cornelissen, G.; Arp, H. P. H.; Pettersen, A.; Hauge, A.; Breedveld, G. D. Assessing PAH and PCB emissions from the relocation of harbour sediments using equilibrium passive samplers. Chemosphere 2008, 72 (10), 1581–1587. (5) Cornelissen, G.; Okkenhaug, G.; Breedveld, G. D.; Sørlie, J.-E. Transport of polycyclic aromatic hydrocarbons and polychlorinated biphenyls in a landfill: A novel equilibrium passive sampler to determine free and total dissolved concentrations in leachate water. J. Hydrol. 2009, 369 (3
4), 253–259. (6) Hawthorne, S. B.; Jonker, M. T. O.; van der Heijden, S. A.; Grabanski, C. B.; Azzolina, N. A.; Miller, D. J. Measuring picogram per liter concentrations of freely dissolved parent and alkyl PAHs (PAH-34),

ARTICLE

using passive sampling with polyoxymethylene. Anal. Chem. 2011, 83 (17), 6754–6761. (7) Cornelissen, G.; Cousins, I. T.; Wiberg, K.; Tysklind, M.; Holmstr€om, H.; Broman, D. Black carbon-dominated PCDD/Fs sorption to soils at a former wood impregnation site. Chemosphere 2008, 72 (10), 1455–1461. (8) McDonough, K. M.; Fairey, J. L.; Lowry, G. V. Adsorption of polychlorinated biphenyls to activated carbon: Equilibrium isotherms and a preliminary assessment of the effect of dissolved organic matter and biofilm loadings. Water Res. 2008, 42 (3), 575–584. (9) Hong, L.; Luthy, R. G. Uptake of PAHs into polyoxymethylene and application to oil-soot (lampblack)-impacted soil samples. Chemosphere 2008, 72 (2), 272–281. (10) ter Laak, T. L.; Busser, F. J. M.; Hermens, J. L. M. Poly(dimethylsiloxane) as passive sampler material for hydrophobic chemicals: Effect of chemical properties and sampler characteristics on partitioning and equilibration times. Anal. Chem. 2008, 80 (10), 3859–3866. (11) Endo, S.; Droge, S. T. J.; Goss, K.-U. Polyparameter linear free energy models for polyacrylate fiber-water partition coefficients to evaluate the efficiency of solid-phase microextraction. Anal. Chem. 2011, 83 (4), 1394–1400. (12) Abraham, M. H.; Ibrahim, A.; Zissimos, A. M. Determination of sets of solute descriptors from chromatographic measurements. J. Chromatogr. A 2004, 1037 (1
2), 29–47. (13) Goss, K.-U. Predicting the equilibrium partitioning of organic compounds using just one linear solvation energy relationship (LSER). Fluid Phase Equilib. 2005, 233 (1), 19–22. (14) Sprunger, L. M.; Acree, W. E., Jr.; Abraham, M. H. Mathematical correlations for gas-to-olive oil, gas-to-saline solution, and saline solution-to-olive oil partition coefficients based on the Goss modified Abraham model. QSAR Comb. Sci. 2008, 27 (7), 890–900. (15) Klamt, A. Conductor-like screening model for real solvents: A new approach to the quantitative calculation of solvation phenomena. J. Phys. Chem. 1995, 99 (7), 2224–2235. (16) Endo, S.; Xu, W.; Goss, K.-U.; Schmidt, T. C. Evaluating coal tar-water partitioning coefficient estimation methods and solute-solvent molecular interactions in tar phase. Chemosphere 2008, 73 (4), 532–538. (17) Arp, H. P. H.; Goss, K.-U. Ambient gas/particle partitioning. 3. Estimating partition coefficients of apolar, polar, and ionizable organic compounds by their molecular structure. Environ. Sci. Technol. 2009, 43 (6), 1923–1929. (18) Goss, K.-U.; Arp, H. P. H.; Bronner, G.; Niederer, C. Nonadditive effects in the partitioning behavior of various aliphatic and aromatic molecules. Environ. Toxicol. Chem. 2009, 28 (1), 52–60. (19) Goss, K.-U. Predicting equilibrium sorption of neutral organic chemicals in various polymers with COSMO-RS. Anal. Chem. 2011, 83 (13), 5304–5308. (20) Hilal, S. H.; Karickhoff, S. W.; Carreira, L. A. Prediction of the vapor pressure boiling point, heat of vaporization and diffusion coefficient of organic compounds. QSAR Comb. Sci. 2003, 22 (6), 565–574. (21) Hilal, S. H.; Saravanaraj, A. N.; Whiteside, T.; Carreira, L. A. Calculating physical properties of organic compounds for environmental modeling from molecular structure. J. Comput.-Aided Mol. Des. 2007, 21 (12), 693–708. (22) Hale, S. E.; Martin, T. J.; Goss, K.-U.; Arp, H. P. H.; Werner, D. Partitioning of organochlorine pesticides from water to polyethylene passive samplers. Environ. Pollut. 2010, 158 (7), 2511. (23) Hale, S. E.; Cornelissen, G.; Arp, H. P. H. Comment on “Partition coefficients of organic contaminants with carbohydrates”. Environ. Sci. Technol. 2010, 45 (3), 1158. (24) Endo, S.; Grathwohl, P.; Schmidt, T. C. Absorption or adsorption? Insights from molecular probes n-alkanes and cycloalkanes into modes of sorption by environmental solid matrices. Environ. Sci. Technol. 2008, 42 (11), 3989–3995. (25) Endo, S.; Grathwohl, P.; Haderlein, S. B.; Schmidt, T. C. Characterization of sorbent properties of soil organic matter and carbonaceous geosorbents using n-alkanes and cycloalkanes as molecular probes. Environ. Sci. Technol. 2009, 43 (2), 393–400. 10131

dx.doi.org/10.1021/es202894k |Environ. Sci. Technol. 2011, 45, 10124–10132

Environmental Science & Technology

ARTICLE

(26) Adams, R. G.; Lohmann, R.; Fernandez, L. A.; MacFarlane, J. K.; Gschwend, P. M. Polyethylene devices: Passive samplers for measuring dissolved hydrophobic organic compounds in aquatic environments. Environ. Sci. Technol. 2007, 41 (4), 1317–1323. (27) Abraham, M. H.; Gola, J. M. R.; Cometto-Muniz, J. E.; Cain, W. S. Solvation properties of refrigerants, and the estimation of their water-solvent and gas-solvent partitions. Fluid Phase Equilib. 2001, 180 (1
2), 41–58. (28) Sprunger, L.; Proctor, A.; Acree, W. E., Jr.; Abraham, M. H. Characterization of the sorption of gaseous and organic solutes onto polydimethyl siloxane solid-phase microextraction surfaces using the Abraham model. J. Chromatogr. A 2007, 1175 (2), 162–173. (29) van Noort, P. C. M.; Haftka, J. J. H.; Parsons, J. R. Updated Abraham solvation parameters for polychlorinated biphenyls. Environ. Sci. Technol. 2010, 44 (18), 7037–7042. (30) Sprunger, L. M.; Achi, S. S.; Acree, W. E., Jr.; Abraham, M. H. Development of correlations for describing solute transfer into acyclic alcohol solvents based on the Abraham model and fragment-specific equation coefficients. Fluid Phase Equilib. 2010, 288 (1
2), 139–144. (31) Goss, K.-U.; Arp, H. P. H.; Bronner, G.; Niederer, C. Partition behavior of hexachlorocyclohexane isomers. J. Chem. Eng. Data 2008, 53 (3), 750–754. (32) Tomaszewski, J. E.; Luthy, R. G. Field deployment of polyethylene devices to measure PCB concentrations in pore water of contaminated sediment. Environ. Sci. Technol. 2008, 42 (16), 6086–6091. (33) Sacks, V. P.; Lohmann, R. Development and use of polyethylene passive samplers to detect triclosans and alkylphenols in an urban estuary. Environ. Sci. Technol. 2011, 45 (6), 2270–2277. (34) Abraham, M. H.; Chadha, H. S.; Whiting, G. S.; Mitchell, R. C. Hydrogen bonding. 32. An analysis of water-octanol and water-alkane partitioning and the Δlog P parameter of seiler. J. Pharm. Sci. 1994, 83 (8), 1085–1100. (35) Abraham, M. H.; Martins, F. Human skin permeation and partition: General linear free-energy relationship analyses. J. Pharm. Sci. 2004, 93 (6), 1508–1523.

10132

dx.doi.org/10.1021/es202894k |Environ. Sci. Technol. 2011, 45, 10124–10132