Article pubs.acs.org/est
Using Polyacrylate-Coated SPME Fibers To Quantify Sorption of Polar and Ionic Organic Contaminants to Dissolved Organic Carbon Joris J.-H. Haftka,* Peter Scherpenisse, Michiel T. O. Jonker, and Joop L. M. Hermens Institute for Risk Assessment Sciences, Utrecht University, Yalelaan 104, 3584 CM Utrecht, The Netherlands S Supporting Information *
ABSTRACT: A passive sampling method using polyacrylate-coated solid-phase microextraction (SPME) fibers was applied to determine sorption of polar and ionic organic contaminants to dissolved organic carbon (DOC). The tested contaminants included pharmaceuticals, industrial chemicals, hormones, and pesticides and represented neutral, anionic, and cationic structures. Prior to the passive sampler application, sorption of the chemicals to the fibers was characterized. This was needed in order to accurately translate concentrations measured in fibers to freely dissolved aqueous concentrations during the sorption tests with DOC. Sorption isotherms of neutral compounds to the fiber were linear, whereas isotherms of basic chemicals covered a nonlinear and a linear range. Sorption of acidic and basic compounds to the fiber was pH-dependent and was dominated by sorption of the neutral sorbate species. Fiber- and DOC-water partition coefficients of neutral compounds were both linearly related to octanol−water partition coefficients (log Kow). The results of this study show that polyacrylate fibers can be used to quantify sorption to DOC of neutral and ionic contaminants, having multiple functional groups and spanning a wide hydrophobicity range (log Kow = 2.5−7.5).
■
INTRODUCTION Synthetic organic contaminants (e.g., pharmaceuticals, industrial products, pesticides, and personal care products) are increasingly emitted into the environment through wastewater, agricultural runoff, and industrial processes.1 These emerging chemicals pose serious problems for the production of safe drinking water.2 Incomplete removal of the compounds during drinking water preparation3 raises questions on the potential health impact of residual contaminants in drinking water. Although individual compounds are considered not to exert acute toxic effects,2,4 limited information is available on the potential (and more subtle) chronic toxic effects of mixtures of contaminants that are present in the ng to μg/L level in natural waters.1 Also, there is little information on the sorption of polar and hydrophilic emerging contaminants1,5 to dissolved organic carbon (DOC). Environmental DOC concentrations have been rising over the past decades, as a result of climate change and decreases in acid rain.6,7 Increasing DOC concentrations in surface waters may cause human health effects through affecting the drinking water preparation cycle. Here, the formation of toxic disinfection byproducts and the levels of contaminants sorbed to DOC entering the cycle may increase.8 Detailed mechanistic insight into the interactions between emerging contaminants and DOC is therefore important in order to be able to predict the mobility of the contaminants in the environment and their fate during the production of drinking water. For neutral contaminants with polar functional groups, sorption to DOC © 2013 American Chemical Society
can be predicted relatively straightforward based on interactions such as hydrogen bond donor/acceptor and van der Waals bonding. For ionic contaminants, the sorption process becomes more complicated however, because in addition to the abovementioned interactions, electrostatic interactions between charged functional groups on DOC and ionic compounds may occur. For example, basic compounds can undergo complexation with negatively charged functional groups, such as carboxylic acids and phenolic groups on DOC, causing the sorption of the neutral species at pH values close to the pKa value of the compound to be exceeded.9 The affinity of negatively charged contaminants for DOC is however less than that of their neutral counterparts, due to repulsive electrostatic interactions.10 Apart from the nonspecific and specific interactions between DOC and neutral or charged species of chemicals, the concentration and chemical characteristics of DOC are also important (i.e., aromaticity, hydrophobic and hydrophilic acids content, and molecular weight).8 Furthermore, solution chemistry variables such as pH and ionic strength are factors to be considered in the sorption of both neutral and ionic contaminants to organic matter.9,11,12 Suitable methods to determine DOC-water partition coefficients (KDOC) of neutral polar contaminants include Received: Revised: Accepted: Published: 4455
January 16, 2013 April 12, 2013 April 15, 2013 April 15, 2013 dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
Figure 1. Molecular structures of the tested compounds. Numbers denote 1. pharmaceuticals, 2. industrial chemicals, 3. hormones, 4. antibacterial and antifungal agents, 5. UV absorbing ingredients in sun screens, 6. herbicides, and 7. pesticides. The asterisks denote atoms that can be ionized. The characters in parentheses denote the different cocktail solutions (labeled A to E) that were used in the determination of KDOC.
fluorescence quenching,13 equilibrium dialysis,14 and passive sampling based on polyacrylate-coated solid-phase microextraction (SPME) fibers.15 DOC-water distribution coefficients (DDOC) for positively charged compounds (i.e., antimicrobial compounds and cationic surfactants) have been determined less frequently, but methods applied for quantification include equilibrium dialysis9,16 and polyacrylate-coated SPME fibers.17 A passive sampler composed of poly(ethylenecovinyl acetate-co-carbon monoxide) has been used to measure KDOC or DDOC of both neutral and cationic compounds.18 Sorption of negatively charged compounds (i.e., antimicrobial compounds) to DOC has been studied with carbowax/ templated polymer resin.19 Among the methods applied, passive sampling polymers, including polyacrylate-coated SPME fibers, have thus been used successfully quite often to measure freely dissolved contaminant concentrations in the presence of DOC. The polyacrylate polymer phase is composed of esterified acrylate of unspecified exact composition and has a high thermal stability and chemical resistance. The phase can undergo nonspecific as well as hydrogen bond interactions with alcoholic, phenolic, and amide groups on sorbates. Polyacrylate is therefore believed to be a stronger hydrogen bond acceptor than polydimethylsiloxane or polyethylene.20,21 This would make polyacrylate an ideal sampling phase for polar functionalized molecules with moderate hydrophobicity (log Kow > 2.5), such as hormones, pharmaceuticals, and pesticides. Currently, only a limited number of studies are available on the sorption of polar and ionic contaminants to DOC. The reasons may be that high DOC concentrations and specific analytical methods are required for studying the relatively weak sorption of hydrophilic compounds. The objective of the
present study was to quantify sorption of a wide range of organic contaminants to 9 natural and 2 commercially available types of DOC. Polyacrylate-coated SPME fibers were applied to measure freely dissolved concentrations of the selection of 13 neutral, 1 anionic, and 4 cationic compounds. The selected compounds with logarithmic octanol to water partition coefficients (log Kow) ranging from 2.5 to 7.5 included hormones, pharmaceuticals, antibacterial and antifungal agents, pesticides, herbicides, UV absorbing ingredients in sun screens, and industrial chemicals. The effect of concentration, pH, and ionic strength on the sorption of the compounds to fibers was studied first to be able to correctly determine their freely dissolved concentrations in the presence of DOC.
■
MATERIALS AND METHODS Chemicals. All tested compounds (purity >98%) were obtained from Dr. Ehrenstorfer (Augsburg, Germany) or Sigma-Aldrich (Zwijndrecht, The Netherlands). The structures of the chemicals are shown in Figure 1, and physicochemical properties are presented in Table S1 in the Supporting Information. Inorganic salts (KH2PO4, K2HPO4, NaN3, NaCl, BaCl2·2H2O, CaCl2·2H2O, AlCl3·6H2O, and Na2SO4) were also obtained from Sigma-Aldrich. Acetonitrile (HPLC-S grade) was obtained from Biosolve (Valkenswaard, The Netherlands). Sorption to Fibers. Glass optical fibers with a 35 μm polyacrylate coating (Vfiber = 15.4 μL/m) were obtained from Polymicro Industries (Phoenix, AZ). Fibers were cut to an appropriate length (varying from 1 to 8 cm), washed three times with pure methanol, and stored in Millipore water until further use. Fiber-water partition coefficients (Kfiber = cfiber/ 4456
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
cwater) were determined by measuring equilibrium concentrations in fiber and water in single solute sorption isotherm experiments using 10 mM CaCl2 (pH = 6.5) or 10 mM potassium phosphate buffer (pH = 5−8) as medium. Equilibrium aqueous concentrations ranged from ∼3 to ∼2500 μg/L for nearly all compounds except for bromophos methyl, clotrimazole, and triclabendazole that were only measured at a single concentration because of the relatively low aqueous solubility of these compounds. All compounds were dissolved in acetonitrile and spiked (V/V = 0.1−0.25%) below aqueous solubility. Systems were closed and equilibrated for 7 to 63 days in the dark on a horizontal shaker at 150 rpm and 21 °C. Equilibrium between fiber and water of the test chemicals was confirmed with kinetic uptake data for the 7 day equilibration time. For all tested chemicals, the polyacrylate fiber length, the dimensions of the system, and the equilibrium times that were used to determine Kfiber are given in Table S2 in the Supporting Information. Depending on the hydrophobicity of the compound, the fiber length and the dimensions of the system were adjusted to prevent excessive depletion of test compounds (>6% of chemical always remained in the water phase, see Table S3). Different fiber lengths did not affect the determination of Kfiber (tested for carbamazepine, atrazine, and testosterone). Upon equilibration, fibers were desorbed in a vial insert in 100 μL of acetonitrile with ibuprofen (1 mg/L) as internal injection standard and shaken with a vortex for approximately 20 s. After allowing the analytes to desorb overnight, the extract was transferred to an autosampler vial, and the fiber was reextracted with 2 × 50 μL of acetonitrile. Repetitive extractions of the fiber showed a negligible amount (5 log units and were linearly related to the log Kow values of the solutes (see Figure 2b), with a maximum deviation of 0.47 log units from the 1:1 line (log Kfiber = (0.93 ± 0.05) × log Kow + (0.22 ± 0.29); r2 = 0.96). This linear trend suggests that equilibrium between fiber and water was reached for all chemicals tested. The linear relationship contrasts previously reported regressions of log Kfiber for polyacrylate versus log Kow, that displayed relatively large scatter.20,29 Reasons for the scatter were assumed inaccuracies in experimentally determined log Kow values and the different ‘sorption’ properties of 1-octanol and polyacrylate,20 reasons that apparently do not, or to a lesser extent, apply to the presently investigated chemicals. Differences in sorption properties of 1-octanol and polyacrylate can be analyzed and interpreted by their respective coefficients in polyparameter linear free energy relationships (pp-LFER). In fact, the coefficient for polarizability (S) has a more important contribution in the pp-LFER of polyacrylate20 compared to that of 1-octanol,30 although the interactions of chemicals between water and 1-octanol or polyacrylate seem rather similar (see Figure 2b). It has been argued that regressions of log Kfiber versus log Kow have a relatively low predictive value as compared to pp-LFER or calculations with COSMOthermX,21 because different intermolecular interactions cannot be taken into account with log Kow models and a large variation of log Kow values exists in the literature. Although we realize that ppLFERs may be more appropriate or powerful, derivation of a pp-LFER is not feasible based on our complete data set, because parameter values are not available for most of our test chemicals. Log Kfiber values could however be estimated with a pp-LFER20 for 8 compounds. The compound descriptors are presented in Table S6 and were taken from refs 31−33. The pp-LFER-predicted log Kfiber value for triclosan was only 0.15 log unit lower than the experimental value. However, the estimated log Kfiber values for carbamazepine, atrazine, bisphenol A, estradiol, and testosterone deviated by −0.78 to +0.29 log units, and even larger deviations were found for triclabendazole and cypermethrin (1.85 and 2.65 log units lower, respectively, than the experimental values). These large differences suggest a model bias or uncertainties in the compound descriptors.31,34 Relative differences in the observed sorption affinity of neutral compounds with different functional groups (e.g., ketone, ether, hydroxyl, or halogen) for the polyacrylate phase can be explained by intermolecular interaction mechanisms. For example, the log Kfiber value of testosterone (2.85 ± 0.02) was 0.89 to 0.94 log units lower than the log Kfiber values of estrone
(version 5.04, San Diego, CA) by weighing with 1/Y2 to aid in fitting the lower concentration range. The pH dependence of the sorption of acidic or basic compounds to the fiber was determined by assuming that the neutral fraction of the compound is the main species that partitions into the fiber coating.19,25,26 The logarithm of the fiber-water distribution coefficient (log Dfiber,acid/base) for acid dissociation of triclosan and estradiol and for base dissociation of chlorpromazine and clotrimazole at each specific pH value was then modeled with eq 3: Dfiber,acid/base = K fiber,neutralfneutral,acid/base
(3)
KDOC (or DDOC) values (L/kg) were calculated as the ratio of the concentration sorbed to DOC, cDOC (μg/kg OC) and the freely dissolved concentration (cfree = cfiber/Kfiber or cfiber/DDOC; μg/L) of the chemicals in the mixed cocktail systems (N = 4; no sorption isotherms were constructed), see eq 4. The concentration sorbed to DOC was calculated here with a mass balance approach that takes the amount sorbed by the fiber into account. The total concentration (ctotal) was determined by directly analyzing the spiking solutions. c KDOC = DOC c free =
■
(ctotalVwater − c freeVwater − c fiberVfiber)/MDOC c free
(4)
RESULTS AND DISCUSSION Sorption of Neutral Organic Compounds to Polyacrylate. Fiber sorption isotherms of the neutral compounds were linear over nearly 2.5 log units (r2 > 0.997, see Figure S1 and Table S3). The log Kfiber values determined for atrazine and estradiol agreed well with literature values: 2.38 measured at 21 °C with 85 μm polyacrylate fibers27 versus 2.28 (this study) for atrazine and 3.66 measured at 22 °C with 7 μm polyacrylate fibers28 versus 3.74 (this study) for estradiol. Competition between neutral compounds for sorption to the fiber was demonstrated to be negligible, based on the similarity of singleand multiple-solute isotherms (see Figure S2). Likewise, the presence of different salts (i.e., 30 mM NaCl, 10 mM CaCl2, 10 mM BaCl2, 5 mM AlCl3, and 10 mM Na2SO4) at the same ionic strength (30 mM) did not affect the log Kfiber values of estradiol (see Figure S3), with estradiol being a representative compound for all neutral compounds. In addition, as expected the log Kfiber values of neutral compounds did not show any pH dependence (see Figure 2a). 4458
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
Figure 3. Concentration dependence of chlorpromazine sorption to polyacrylate-coated fibers at pH 5 and 7 (a; the arrow indicates the spiked concentration in systems containing DOC) and pH dependence of sorption of ionic organic compounds to the fibers (b; charge of the basic and acidic compounds is given in brackets). Dotted and dashed lines represent linear and nonlinear contributions, respectively, to the sorption isotherms.
fiber (see Figures S4 and S5). Because of the observed competition effects, the final sorption experiments with DOC were performed with only one basic compound in a mixture together with neutral compounds (see Figure 1). The presence of different salts did not influence the log Dfiber values of chlorpromazine, as shown in Figure S3. Furthermore, increased concentrations of potassium phosphate buffer (0.1−100 mM) and sodium chloride (5−500 mM) did not affect the sorption of chlorpromazine to the fiber (data not shown). Sorption of chlorpromazine to the fiber was however observed to be strongly dependent on the pH of the medium. The log Dfiber values of the chemical increased by 1.74 log units, going from pH 5 (log Dfiber of 1.38) to pH 7 (3.12) at an aqueous equilibrium concentration of 500 μg/L (see Figure 3a). Also, the maximum sorption capacity of polyacrylate for sorption of positively charged chlorpromazine increased going from pH 5 to 7, most probably due to deprotonation of the negatively charged sites on the surface of polyacrylate at higher pH values (see eq 1).17 Note that both nonlinear and linear sorption contributions decrease at lower pH and the relative contribution of linear sorption even becomes higher at pH 5, as compared to pH 7 (see Table S7). The pH dependence of sorption of all charged chemicals is shown in Figure 3b and will be discussed below. At pH 7 and an aqueous equilibrium concentration of 500 μg/L, sorption of promazine to polyacrylate was about 0.75 log units lower than sorption of chlorpromazine and triflupromazine. The only difference in the molecular structures of these compounds is the substitution of a hydrogen atom in promazine by chlorine and a carbon trifluorine group in chlorpromazine and triflupromazine, respectively (see Figure 1 and Table S1). These changes in sorption strength in response to changes in the molecular structure suggest that sorption to the fiber is mainly driven by the hydrophobicity of the relatively small fraction of neutral compound present. The log Dfiber values calculated for promazine, chlorpromazine, and triflupromazine at pH 7 also fit the linear relationship between log Dfiber and log Dow nicely (see Figure 2b). This implies that sorption of basic compounds to polyacrylate at pH 7 can be predicted on the basis of the pH-dependent distribution in 1-octanol, which assumes that predominantly the neutral fraction partitions to both 1-octanol and polyacrylate. A larger deviation from the 1:1 line in Figure 2b was found for the acidic compound triclosan (log Dfiber = 5.58 ± 0.02; f neutral = 0.936). This chemical sorbed very strongly to the fiber, which may be explained by the highly polarizing chlorine atoms that are directly bound to the
(3.79 ± 0.03) and estradiol (3.74 ± 0.02). The main interactions between the hormones and the polyacrylate phase are hydrogen bond donor interactions (hydroxyl groups), hydrogen bond acceptor interactions (hydroxyl and ketone groups), and van der Waals interactions. The conjugated ring in estrone and estradiol may be responsible for the additional interactions with polyacrylate, explaining the difference between log Kfiber values for estradiol/estrone and testosterone. Other clear differences in log Kfiber values were observed for the synthetic pyrethroids cypermethrin, deltamethrin, and permethrin. Log Kfiber increased from cypermethrin (6.71 ± 0.11) to deltamethrin (6.83 ± 0.08), which will simply be caused by the substitution of chlorine by the larger and more hydrophobic bromine atom. Subsequent loss of the carbonitrile group, going from cypermethrin to permethrin will also induce more favorable van der Waals and hydrogen bond interactions between permethrin and polyacrylate. Sorption of Ionic Organic Compounds to Polyacrylate. Fiber-water distribution was quantified for the ionic chemical chlorpromazine over a wide concentration range (∼100 μg/L to ∼100 mg/L). Chlorpromazine is a basic organic compound that is positively charged at pH 7 ( f neutral = 0.005; pKa = 9.3). Sorption of this chemical to polyacrylate resulted in a nonlinear sorption isotherm with a Freundlich exponent (n) of 0.75 − 0.79 at low concentrations (see Figures 3a and S4). At higher aqueous concentrations, the sorption isotherm became however linear again (n = 0.96−0.99 for the three highest concentrations). This phenomenon can be explained by a combination of nonlinear adsorption of the basic compound to negatively charged surface sites on the fiber17 and linear absorption of the neutral compound into the polyacrylate phase of the fiber. The sorption isotherms for chlorpromazine at pH 5 and 7 were fit accordingly with a dual mode model that incorporated base dissociation of chlorpromazine (see eqs 1 and 2). The resulting fitting parameters are shown in Table S7. The maximum amount of positively charged chlorpromazine that can sorb to negatively charged sorption sites on the fiber is 7.37 mmol/L (Cs,max in eq 1). Interestingly, the calculated linear sorption contribution still represents 32% of the total sorbed amount at pH 7 (at ctotal of 500 μg/L), even though the percentage of the neutral species of chlorpromazine is only 0.47%. This suggests that at pH 7 the neutral species has a much higher affinity for polyacrylate than the charged species of chlorpromazine. During multiple-solute experiments, the addition of a mixture of basic compounds (promazine, chlorpromazine, and triflupromazine) resulted in competition for sorption sites on the 4459
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
because of a different protonation state of the DOC due to the higher pH value. The log KDOC values for the DOC sample from Bangor covered a range of 2.7 log units and were linearly related to log Kow values (see Figure 4) according to log KDOC = (0.62 ±
benzene ring, in addition to the hydrogen bond donor/ accepting interactions of the hydroxyl group. For each pH unit of increase, log Dfiber values increased with about 0.9 log unit for promazine, chlorpromazine, and triflupromazine (see Figure 3b). The log Dfiber values of chlorpromazine and clotrimazole became constant at a pH of ∼2 log units above the pKa value, which indicates that at these high pH values only partitioning of the neutral compound to the fiber occurred. For the acidic compounds triclosan and estradiol, constant log Dfiber values were obtained after the compounds were fully protonated at a pH of ∼2 log units below the pKa value. This effect has previously also been observed for sorption of organic acids and bases to polyacrylate,25 sorption of quinolones by carbowax/template polymer resin,19 and sorption of alkyphenol and triclosans to polyethylene sheets.26 Using the parameters resulting from fitting eq 1 to the experimental data at pH 7, it was estimated that the linear sorption contribution was 99% at the pKa value of chlorpromazine (pH = 9.3; f neutral = 50%). The nonlinear sorption contribution of chlorpromazine therefore becomes negligible at higher pH values and is superseded by the linear sorption contribution of the neutral fraction. Interestingly, by applying eq 3, the data from Figure 3b can be used to estimate the pKa values of triclosan, clotrimazole, chlorpromazine, and estradiol. The resulting pKa values were 6.11 and 8.92 for the basic compounds clotrimazole and chlorpromazine, respectively, which are close to the literature values (being 5.9935 and 9.2,36 respectively). Similarly, the pKa values of triclosan (pKa = 8.00) and estradiol (pKa = 10.58) were close to the literature value of 8.137 and an estimated value of 10.4,38 respectively (see Table S1). Sorption of Neutral Organic Compounds to DOC. The log KDOC values of the tested neutral compounds were calculated using a mass balance approach (see eq 4). The resulting log KDOC values of the less hydrophobic compounds had relatively large errors. All data were therefore scrutinized to obtain a reliable data set of log KDOC values. First, the concentrations in fibers from experiments with and without DOC were compared with an unpaired Student’s t test (assuming equal variances; N = 4). This was done to confirm that the decrease in cfiber due to sorption of the test compounds to DOC was statistically significant. Second, the relative standard deviation (RSD) of the amount sorbed to DOC was derived, using standard error propagation calculations, and assuming that the RSD value was only caused by variations in the measured cfiber and Kfiber values. For all cases for which statistically significant (P < 0.05) differences were observed, RSD values of the mass sorbed to DOC were lower than 20 to 30% when the mass sorbed to DOC was more than 30% of the total mass in the system. Therefore, log KDOC values were considered reliable only when the mass of test compound sorbed to DOC was >30%. The result of this quality assurance step however was that out of the 11 DOC samples tested in total (with variable pH and DOC concentrations), only one DOC sample (Bangor, UK, having an organic carbon concentration of 282 mg C/L and a pH of 5.38) allowed the calculation of log KDOC values for nearly all neutral compounds. Exceptions were carbamazepine, atrazine, estrone, and bisphenol A, for which no data could be reported (see Table S8). A DOC sample with a similar organic carbon concentration (Rottighe Meenthe 2, NL; 288 mg C/L; pH = 7.13) resulted in fewer data for the more hydrophilic compounds, probably
Figure 4. Logarithmic DOC-water partition/distribution coefficients of neutral and ionic organic compounds versus logarithmic octanol to water partition/distribution coefficients at pH 7. The DOC sample was from Bangor, UK. The linear free energy relationship (dotted line) for sorption of polycyclic aromatic hydrocarbons to natural DOC is taken from ref 39.
0.03) × log Kow + (0.92 ± 0.19); r2 = 0.98. This relationship derived for neutral chemicals is less steep than an LFER derived from literature data on polycyclic aromatic hydrocarbon (PAH) sorption to natural DOC: log KDOC = (1.18 ± 0.13) × log Kow − (1.56 ± 0.72); r2 = 0.58.39 Although the coefficients in this type of LFER may vary depending on the type of DOC studied and the study from Burkhard et al.39 did not include DOC from Bangor, the observed slopes and intercepts suggest that the sorption mechanism for neutral polar chemicals is likely to be quite different from that for PAHs. Note that the same limitations mentioned above for the LFER based on log Kow for polyacrylate also apply here in the case of DOC. For the DOC sample from Bangor, similar log KDOC values were calculated for testosterone (3.18 ± 0.08 L/kg) and estradiol (3.23 ± 0.07 L/kg), suggesting similar interaction mechanisms between these compounds and DOC, in contrast to what was observed for polyacrylate (see above). For octocrylene, reliable log KDOC values could be calculated for nearly all DOC samples (N = 10). These ranged from 4.78 ± 0.11 to 5.64 ± 0.08 L/kg and thus showed a large variation of nearly 1 log unit. Similarly, differences in DOC to air partition coefficients of more than 1 log unit have previously been observed for sorption of 80100 nonpolar and polar chemicals to aquatic and terrestrial humic and fulvic acids.40 For the KDOC values of octocrylene in the present study, no significant relationship could be found with DOC properties that had been determined (pH, SUVA, hydrophilic/hydrophobic acid and bases). Multiple linear regression with all DOC properties also did not result in any significant correlations (data not shown). Other components or properties of DOC are therefore probably responsible for the observed variation in the log KDOC values. KDOC determinations at equal pH and DOC concentrations may gain more insight into the relevant sorption properties of DOC samples from different origins. Sorption of Ionic Organic Compounds to DOC. The DDOC values for basic compounds were calculated with pHdependent Dfiber values for promazine, chlorpromazine, and triflupromazine that were derived with linear regression (log Dfiber = a × pH + b). The sorption isotherms of these 4460
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
more research is required to exactly quantify the above effects and their possible interplay. The presence of (increasing concentrations of) DOC can substantially affect the fate and transport of many compounds in surface waters and soils, including the presently studied basic compounds. For other chemicals, such as small neutral compounds (e.g., chlorinated and brominated methanes; disinfection byproducts), DOC may however not be an important sorption phase. No appreciable sorption of these compounds was for instance observed in the presence of exceptionally high concentrations (10,000−12,000 mg/L) of humic and fulvic acids.41 For the testing compounds in this study, the lowest log KDOC value was measured for testosterone (3.18 ± 0.08 L/kg), which implies only 1.5% sorption to DOC at an environmentally relevant concentration of 10 mg C/L. Sorption of this type of hydrophilic compounds to DOC can however only be quantified with polyacrylate-coated fibers at relatively high DOC concentrations as has been demonstrated in this study.
compounds were linear at the concentrations applied during the experiments with DOC (see arrow in Figure 3a). The pHdependent Dfiber values for clotrimazole were calculated with eq 3. The DDOC values were subsequently calculated with eq 4 using these pH-dependent Dfiber values. The percentage sorption to DOC was well above 30% for the basic compounds chlorpromazine (log DDOC = 3.68−4.78 L/kg; N = 8), triflupromazine (log DDOC = 3.68−4.76 L/kg; N = 11), and clotrimazole (log DDOC = 4.12−5.14 L/kg; N = 11). Because also cfiber of the systems containing DOC were significantly different (P < 0.05) from those without DOC and the RSD values of the calculated mass sorbed to DOC were all below 20%, these log DDOC values should be considered reliable. The log DDOC values for all basic compounds that passed the quality criteria are presented in Table S9. The strong sorption of chlorpromazine, triflupromazine, and clotrimazole to DOC is probably caused by specific adsorption of the ionized nitrogen atoms to deprotonated carboxylic acid and phenolic groups on DOC, which will occur in addition to van der Waals interactions of the neutral parts of the chemicals to uncharged moieties of DOC. The log DDOC values for chlorpromazine and triflupromazine do not follow the relationship of log DDOC versus log Dow observed for the other compounds (see Figure 4), because these compounds are completely charged. As expected, octanol therefore is not a suitable predictor of organic base sorption to DOC. Still, for clotrimazole, for which the neutral fraction at pH 5.38 for DOC from Bangor is only 0.2, the log DDOC value is close to the general relationship, which suggests that the neutral form of this compound is the dominating sorption species. The log DDOC values for promazine, chlorpromazine, and triflupromazine displayed relatively high standard deviations as compared to values determined for the other neutral compounds. Most probably this is due to the fact that the pH of the DOC samples was not buffered, which might have caused small changes in the pH that directly will have affected the freely dissolved concentrations of these compounds. Similar to what was observed for the neutral compounds, no significant relationships were found between the log DDOC values of basic compounds and DOC properties. Sorption of basic compounds to DOC is therefore (also) a complex process, all the more since it will generally be dependent on several additional factors, including the pH9,11 and the ionic strength of the medium,12 and the solute concentration.16 For instance, sorption of basic compounds will generally increase with increasing pH, because carboxylic and phenolic acid groups on DOC are gradually deprotonated. The sorption affinity will therefore show a maximum near the pKa of the compound.9,11 Above the pKa, sorption will decrease again, because the compound becomes neutral. Furthermore, sorption of basic chemicals will decrease with increasing ionic strength of the medium, because of competition for sorption sites on DOC with inorganic cations (e.g., Na+, K+, Ca2+, Mg2+).12 Also, sorption of the compounds will increase with decreasing freely dissolved concentrations, because of the nonlinear sorption to a limited number of (heterogeneous) sorption sites on DOC.16 Finally, van der Waals interactions and hydrogen bond/donor interactions will play an important role in determining the hydrophobic sorption contribution to DOC. Knowledge of the effects of all these factors is crucial for the prediction of the sorption affinity of basic compounds. The present log DDOC values should therefore be considered operational ones, and
■
ASSOCIATED CONTENT
S Supporting Information *
Full description of Materials and Methods and additional figures and tables. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This project was funded by the Dutch Ministry of Environment and Infrastructure under the ERANET Environment and Health program entitled “Environmental change and rising DOC trends: Implications for public health”. David Hughes and Tim Jones (Bangor University, UK) are acknowledged for performing the fractionations of the DOC samples and for sampling DOC from a soil core, respectively. Jarkko Akkanen (University of Eastern Finland) is acknowledged for sampling DOC from Finnish lakes. Stephan van der Heijden (IRAS, Utrecht University) is acknowledged for preparing the DOC samples.
■
REFERENCES
(1) Schwarzenbach, R. P.; Egli, T.; Hofstetter, T. B.; von Gunten, U.; Wehrli, B. Global water pollution and human health. Annu. Rev. Environ. Resour. 2010, 35, 109−136. (2) Houtman, C. J. Emerging contaminants in surface waters and their relevance for the production of drinking water in Europe. J. Integr. Environ. Sci. 2010, 7 (4), 271−295. (3) Ternes, T. A.; Meisenheimer, M.; McDowell, D.; Sacher, F.; Brauch, H.-G.; Haist-Gulde, B.; Preuss, G.; Wilme, U.; Zulei-Seibert, N. Removal of pharmaceuticals during drinking water treatment. Environ. Sci. Technol. 2002, 36 (17), 3855−3863. (4) Schriks, M.; Heringa, M. B.; van der Kooi, M. M. E.; de Voogt, P.; van Wezel, A. P. Toxicological relevance of emerging contaminants for drinking water quality. Water Res. 2010, 44 (2), 461−476. (5) Giger, W. Hydrophilic and amphiphilic water pollutants: using advanced analytical methods for classic and emerging contaminants. Anal. Bioanal. Chem. 2009, 393 (1), 37−44.
4461
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
Environmental Science & Technology
Article
dissolved organic matter. Environ. Sci. Technol. 2009, 43 (5), 1379− 1385. (25) Escher, B. I.; Berg, M.; Mühlemann, J.; Schwarz, M. A. A.; Hermens, J. L. M.; Vaes, W. H. J.; Schwarzenbach, R. P. Determination of liposome/water partition coefficients of organic acids and bases by solid-phase microextraction. Analyst 2002, 127, 42−48. (26) 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. (27) Valor, I.; Pérez, M.; Cortada, C.; Apraiz, D.; Moltó, J. C.; Font, G. SPME of 52 pesticides and polychlorinated biphenyls: extraction efficiencies of the SPME coatings poly(dimethylsiloxane), polyacrylate, poly(dimethylsiloxane)-divinylbenzene, carboxen-poly(dimethylsiloxane), and carbowax-divinylbenzene. J. Sep. Sci. 2001, 24 (1), 39−48. (28) Heringa, M. B.; Pastor, D.; Algra, J.; Vaes, W. H. J.; Hermens, J. L. M. Negligible depletion solid-phase microextraction with radiolabeled analytes to study free concentrations and protein binding: an example with [3H]estradiol. Anal. Chem. 2002, 74 (23), 5993−5997. (29) Verbruggen, E. M. J.; Vaes, W. H. J.; Parkerton, T. F.; Hermens, J. L. M. Polyacrylate-coated SPME fibers as a tool to simulate body residues and target concentrations of complex organic mixtures for estimation of baseline toxicity. Environ. Sci. Technol. 2000, 34 (2), 324−331. (30) Abraham, M. H.; Zhao, Y. H. Characterisation of the water/onitrophenyl octyl ether system in terms of the partition of nonelectrolytes and of ions. Phys. Chem. Chem. Phys. 2005, 7 (12), 2418−2422. (31) Tülp, H. C.; Goss, K.-U.; Schwarzenbach, R. P.; Fenner, K. Experimental determination of LSER parameters for a set of 76 diverse pesticides and pharmaceuticals. Environ. Sci. Technol. 2008, 42 (6), 2034−2040. (32) Zissimos, A. M.; Abraham, M. H.; Du, C. M.; Valko, K.; Bevan, C.; Reynolds, D.; Wood, J.; Tam, K. Y. Calculation of Abraham descriptors from experimental data from seven HPLC systems; evaluation of five different methods of calculation. J. Chem. Soc., Perkin Trans. 2 2002, 12, 2001−2010. (33) Abraham, M. H.; Ibrahim, A.; Zhao, Y.; Acree, W. E., Jr. A data base for partition of volatile organic compounds and drugs from blood/plasma/serum to brain, and an LFER analysis of the data. J. Pharm. Sci. 2006, 95 (10), 2091−2100. (34) Bronner, G.; Goss, K.-U. Predicting sorption of pesticides and other multifunctional organic chemicals to soil organic carbon. Environ. Sci. Technol. 2011, 45 (4), 1313−1319. (35) Shalaeva, M.; Kenseth, J.; Lombardo, F.; Bastin, A. Measurement of dissociation constants (pKa values) of organic compounds by multiplexed capillary electrophoresis using aqueous and cosolvent buffers. J. Pharm. Sci. 2008, 97 (7), 2581−2606. (36) Syrres Web site, PHYSPROP database. http://www.syrres.com/ what-we-do/databaseforms.aspx?id=386 (accessed April 23, 2013). (37) Tixier, C.; Singer, H. P.; Canonica, S.; Müller, S. R. Phototransformation of triclosan in surface waters: a relevant elimination process for this widely used biocide - laboratory studies, field measurements, and modeling. Environ. Sci. Technol. 2002, 36 (16), 3482−3489. (38) SPARC Web site. http://archemcalc.com/sparc (accessed April 23, 2013). (39) Burkhard, L. P. Estimating dissolved organic carbon coefficients for nonionic organic chemicals. Environ. Sci. Technol. 2000, 34 (22), 4663−4668. (40) Niederer, C.; Schwarzenbach, R. P.; Goss, K.-U. Elucidating differences in the sorption properties of 10 humic and fulvic acids for polar and nonpolar organic chemicals. Environ. Sci. Technol. 2007, 41 (19), 6711−6717. (41) Neale, P. A.; Escher, B. I.; Goss, K.-U.; Endo, S. Evaluating dissolved organic carbon-water partitioning using polyparameter linear free energy relationships: Implications for the fate of disinfection byproducts. Water Res. 2012, 46 (11), 3637−3645.
(6) Freeman, C.; Evans, C. D.; Monteith, D. T.; Reynolds, B.; Fenner, N. Export of organic carbon from peat soils. Nature 2001, 412 (6849), 785. (7) Evans, C. D.; Monteith, D. T.; Cooper, D. M. Long-term increases in surface water dissolved organic carbon: observations, possible causes and environmental impacts. Environ. Pollut. 2005, 137 (1), 55−71. (8) Matilainen, A.; Gjessing, E. T.; Lahtinen, T.; Hed, L.; Bhatnagar, A.; Sillanpää, M. An overview of the methods used in the characterisation of natural organic matter (NOM) in relation to drinking water treatment. Chemosphere 2011, 83 (11), 1431−1442. (9) Sibley, S. D.; Pedersen, J. A. Interaction of the macrolide antimicrobial clarithromycin with dissolved humic acid. Environ. Sci. Technol. 2008, 42 (2), 422−428. (10) Tülp, H. C.; Fenner, K.; Schwarzenbach, R. P.; Goss, K.-U. pHdependent sorption of acidic organic chemicals to soil organic matter. Environ. Sci. Technol. 2009, 43 (24), 9189−9195. (11) Karthikeyan, K. G.; Chorover, J. Humic acid complexation of basic and neutral polycyclic aromatic compounds. Chemosphere 2002, 48 (9), 955−964. (12) Droge, S.; Goss, K.-U. Effect of sodium and calcium cations on the ion-exchange affinity of organic cations for soil organic matter. Environ. Sci. Technol. 2012, 46 (11), 5894−5901. (13) Yamamoto, H.; Liljestrand, H. M.; Shimizu, Y.; Morita, M. Effects of physical-chemical characteristics on the sorption of selected endocrine disruptors by dissolved organic matter surrogates. Environ. Sci. Technol. 2003, 37 (12), 2646−2657. (14) Stumpe, B.; Marschner, B. Dissolved organic carbon from sewage sludge and manure can affect estrogen sorption and mineralization in soils. Environ. Pollut. 2010, 158 (1), 148−154. (15) Neale, P. A.; Escher, B. I.; Schäfer, A. I. Quantification of solutesolute interactions using negligible-depletion solid-phase microextraction: measuring the affinity of estradiol to bulk organic matter. Environ. Sci. Technol. 2008, 42 (8), 2886−2892. (16) Richter, M. K.; Sander, M.; Krauss, M.; Christl, I.; Dahinden, M. G.; Schneider, M. K.; Schwarzenbach, R. P. Cation binding of antimicrobial sulfathiazole to Leonardite humic acid. Environ. Sci. Technol. 2009, 43 (17), 6632−6638. (17) Chen, Y.; Droge, S. T. J.; Hermens, J. L. M. Analyzing freely dissolved concentrations of cationic surfactant utilizing ion-exchange capability of polyacrylate coated solid-phase microextraction fibers. J. Chromatogr., A 2012, 1252, 15−22. (18) Magnér, J. A.; Alsberg, T. E.; Broman, D. The ability of a novel sorptive polymer to determine the freely dissolved fraction of polar organic compounds in the presence of fulvic acid or sediment. Anal. Bioanal. Chem. 2009, 395 (5), 1525−1532. (19) Holten Lützhøft, H. C.; Vaes, W. H. J.; Freidig, A. P.; HallingSørensen, B.; Hermens, J. L. M. Influence of pH and other modifying factors on the distribution behavior of 4-quinolones to solid phases and humic acids studied by ″negligible-depletion″ SPME-HPLC. Environ. Sci. Technol. 2000, 34 (23), 4989−4994. (20) 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. (21) Endo, S.; Hale, S. E.; Goss, K.-U.; Arp, H. P. H. Equilibrium partition coefficients of diverse polar and nonpolar organic compounds to polyoxymethylene (POM) passive sampling devices. Environ. Sci. Technol. 2011, 45 (23), 10124−10132. (22) Weishaar, J. L.; Aiken, G. R.; Bergamaschi, B. A.; Fram, M. S.; Fujii, R.; Mopper, K. Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environ. Sci. Technol. 2003, 37 (20), 4702−4708. (23) Kopinke, F.-D.; Ramus, K.; Poerschmann, J.; Georgi, A. Kinetics of desorption of organic compounds from dissolved organic matter. Environ. Sci. Technol. 2011, 45 (23), 10013−10019. (24) Ter Laak, T. L.; Van Eijkeren, J. C. H.; Busser, F. J. M.; Van Leeuwen, H. P.; Hermens, J. L. M. Facilitated transport of polychlorinated biphenyls and polybrominated diphenyl ethers by 4462
dx.doi.org/10.1021/es400236a | Environ. Sci. Technol. 2013, 47, 4455−4462
