Sediment-Associated Reactions of Aromatic Amines. 1. Elucidation of

Sarah G. Pati , Kwanghee Shin , Marita Skarpeli-Liati , Jakov Bolotin , Soren N. Eustis , Jim C. Spain , and Thomas B. Hofstetter. Environmental Scien...
0 downloads 0 Views 84KB Size
Environ. Sci. Technol. 2001, 35, 2470-2475

Sediment-Associated Reactions of Aromatic Amines. 1. Elucidation of Sorption Mechanisms E R I C J . W E B E R , * ,† DALIZZA COLO Ä N,† AND GEORGE L. BAUGHMAN‡ U.S. Environmental Protection Agency, National Exposure Research Laboratory, 960 College Station Road, Athens, Georgia 30605-2720, and Department of Textiles, Merchandising and Interiors, University of Georgia, Athens, Georgia 30613

Sorption of aromatic amines to sediments and soils can occur by both reversible physical processes and irreversible chemical processes. To elucidate the significance of these sorption pathways, the sorption kinetics of aniline and pyridine were studied in resaturated pond sediment. Aniline and pyridine behaved quite differently in the sedimentwater systems. The sorption kinetics of pyridine were quite fast, reaching equilibrium within 1-2 h. In contrast, the sorption kinetics of aniline were characterized by a rapid initial loss of aniline from the aqueous phase followed by a much slower rate of disappearance. The rapid initial sorption of aniline upon respiking after an equilibration period of 200 h, and results of sorption kinetic studies as a function of substrate concentration, demonstrated that sorptive sites were not being saturated at the nominal concentration of aniline. Sequential extraction of a sediment treated with 14C-labeled pyridine and aniline suggested that pyridine was bound primarily through a reversible cation-exchange process, whereas aniline sorbed through both cation-exchange and covalent binding processes. At longer reaction periods sorption became increasingly dominated by covalent binding. The reaction kinetics for the slow, irreversible sorption of aniline appeared to be limited by the reactivity and/or availability of covalent binding sites. The initial rate and extent of aniline sorption was pH dependent (sorption increased with decreasing pH). At pH values above the pKa of aniline, sorption kinetics for the slower, irreversible loss of aniline were independent of pH.

Introduction Aromatic amines comprise an important class of environmental contaminants. The aromatic amines are building blocks for many textile dyes, agrochemicals, and other classes of synthetic chemicals. Concern exists over the loss of the aromatic amines to the environment during production processes or improper treatment of industrial waste streams. In addition, aromatic amines can enter the environment from the reduction of azo dyes, polynitroaromatic munitions (e.g., TNT), dinitroherbicides, and hydrolytic degradation of * Corresponding author phone: (706)355-8224; fax: (706)355-8202; e-mail: [email protected]. † U.S Environmental Protection Agency. ‡ University of Georgia. 2470

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 12, 2001

numerous agrochemicals (1). A thorough understanding of how these chemicals interact with natural surfaces is important to successfully model their transport and transformation in aquatic and soil ecosystems. The physical and chemical properties of aromatic amines suggest that a number of sorption reactions must be considered for this class of compounds. Because aromatic amines are Lewis bases, the potential exists for the formation of the protonated species in natural aquatic ecosystems suggesting the possibility for sorption to occur through cation exchange. Although many aromatic amines have significant water solubility, hydrophobic partitioning could potentially be the dominant sorption process for aromatic amines with multiple aromatic rings or hydrophobic substituents (e.g., alkyl groups). In addition to sorption to sediments by reversible physical processes, the reactivity of the amino functional group requires consideration of chemical processes that result in formation of covalent bonds with constituents of the sediment matrix. Covalent binding has been proposed to result from the addition of the nucleophilic amino functional group to electrophilic sites (i.e., carbonyl moieties) and/or oxidative mechanisms resulting in the formation of radical species that couple with sediment-bound radicals (2-10). 15NMR studies have provided direct spectroscopic evidence for the reaction pathway occurring through nucleophilic addition to carbonyl moieties in dissolved organic matter (11, 12). These studies demonstrated that 15N-labeled aniline became incorporated in the form of anilinohydroquinone, anilinoquinone, anilide, imine, and heterocyclic nitrogen, the latter comprising 50% or more of the bound amine. The formation of heterocyclic nitrogen suggests the importance of the reaction with carbonyl groups such as ketones, 1,4-diketones, and β-dicarbonyls, including β-diketones. The complexity of the sorption processes for aromatic amines often manifests itself in the reaction kinetics observed for the sorption of these chemicals to soils and sediments. The sorption of aromatic amines to these surfaces is often described by biphasic kinetics, which has generally been attributed to a change in sorption mechanism with time (2, 6, 9, 13). Sorption is proposed to occur initially by a rapid, reversible equilibrium, followed by a slower, irreversible sorption process (14). The rapid sorption step has been attributed to electrostatic interactions, hydrophobic partitioning, and the formation of labile amine-carbonyl adducts (e.g., imines) (4, 15). The slower process has been attributed to irreversible covalent binding (4, 9, 10, 16). Limitations in the reactivity and/or availability of sorption sites can also contribute to the occurrence of biphasic sorption kinetics. Toward the goal of providing further insight into processes controlling the environmental fate of aromatic amines, we report on the initial phase of our work in which we have studied the sediment-associated reactions of aniline and pyridine in resaturated pond sediment. Selected chemical and physical properties of aniline and pyridine are summarized in Table 1. Pyridine is a nitrogen-heterocyclic compound (NHC). The NHCs are an important class of environmental contaminants common to waste streams generated from energy development technologies such as coal gasification and shale oil extraction. Like aniline, pyridine is a basic compound, thus the potential for formation of the protonated species exists. In contrast to aniline, however, the nitrogen in pyridine is part of an aromatic ring, thus precluding covalent binding to sediments and soils. In this paper we present results of our laboratory studies that compare the sorption reactions of aniline and pyridine 10.1021/es001759d CCC: $20.00

 2001 American Chemical Society Published on Web 05/04/2001

TABLE 1. Selected Chemical and Physical Properties of Aniline and Pyridine

a

Reference 29.

b

Reference 30. c Reference 31.

in a sediment-water system. Sequential extractions were performed to differentiate between reversible and irreversible sorption processes controlling the interaction of aniline and pyridine with sediment. We have also investigated the effect of system variables such as substrate concentration and solution pH on sorption kinetics. Subsequent manuscripts will report on the effect of substituent groups on the sorption pathways and kinetics of the aromatic amino group in sediment-water systems (QSAR analysis) and the development of predictive kinetic models.

Experimental Section Chemicals. Aniline (99% pure, Aldrich) was vacuum distilled prior to use. Pyridine (95% pure; Aldrich) was used as received. Ring labeled 14C-aniline (Sigma, specific activity 58.6 mCi/ mmol) and 14C-pyridine (Sigma, specific activity 15.5 mCi/ mmol) were used as received. All solvents used were of high purity (Burdick and Jackson). Sediment. Sediment was collected from Cherokee Park pond in Athens, GA. The sediment and associated water were collected by scooping up the first 5-10 cm of the sediment surface in 1-L glass jars at a depth of 30-60 cm below the water surface. The jars were capped after being brought to the surface. The samples then were transported to the laboratory and passed through a 1-mm sieve. The sediment was air-dried by spreading to a depth of several mm in glass pans and allowing to stand for 5 days. The dried sediment was homogenized and passed through a 1-mm sieve. The percent organic carbon of the Cherokee Park sediment was 3.3 ( 0.5. The percentages of sand, silt, and clay in the sediment were 36.5, 32.1, and 31.4%, respectively. The CEC for Cherokee Park sediment was 11.8 mequiv/100 g, the base saturation was 3.98%, and the particle density was 2.41 g/cm3. The clay mineralogy of Cherokee Park sediment was dominated by kaolinite. Other minerals detected in smaller amounts included chlorite/vermiculite, hydrous mica, gibbsite, goethite, and quartz. Analytical Methods. The LC system consisted of two Gilson 305/302 gradient chromatographic pumps, an Applied Biosystems 783 programmable wavelength detector, a Rheodyne 7161 injector (200-µL sample loop), an Alcott 728 autosampler, and a Hewlett-Packard 3396A integrator. The detection wavelength was 235 nm. The analytical column was an Alltech RSIL C18 (25 cm long × 4.6 mm i.d., 5-µm particle size). The columns were protected with an AlltechApplied Science Adsorbosphere C18 cartridge guard column. The eluent consisted of 20% acetonitrile in water buffered at pH 7 with 5-mM phosphate buffer. Kinetic Studies. The air-dried pond sediment, double distilled water, and a 250-mL Erlenmeyer flask were heat sterilized (20 min, 121 °C, 20 psi). Using sterile techniques, 97 mL of H2O and 1.0 mL of the appropriate concentration of an aqueous stock solution of aniline, which had been sterilized by filtering through a 0.2-µm Nylon 66 filter, were added to 5.0 g of sediment in the Erlenmeyer flask. The flask

was placed on a temperature-controlled platform shaker at 20 °C. At preselected times, a 3-mL aliquot was removed while stirring using sterile techniques. The aliquot was transferred to two 1.5-mL polypropylene centrifuge tubes and centrifuged for 10 min at 14 000 rpm (Eppendorf 5415 C microcentrifuge). The supernatant was removed and transferred to glass autosampler vials for LC analysis. For kinetic studies in which the solution pH was altered, buffer solution (filter sterilized) was added to 5.0 g of sediment in 87-97 mL of H2O. The buffer components and concentrations were as follows: pH 3.75 (10 mL of 0.033 M citric acid/0.17 M citrate), pH 5.27 (1.0 mL of 0.75 M NH4PO4H2/ 0.25 M NH4PO4Na2), pH 6.14 (1.0 mL of 0.40 M NH4PO4H2/ 0.60 M NH4PO4Na2), and pH 7.37 (1.0 mL of 0.05 M NH4PO4H2/0.95 M NH4PO4Na2). Buffer selections kept ionic strength variations within a minimum. The buffered sediment slurries were equilibrated for 24 h. Solution pH was measured just prior to the addition of substrate. Sequential Extractions. One gram of air-dried pond sediment and 9.0 mL of distilled water were added to a series of 15-mL test tube. After rotating the tubes end-over-end for 24 h, the tubes were spiked with 50 µL of a 1 mM solution of 14C-aniline hydrochloride or 14C-pyridine to give a final concentration of 5 nCi/mL of 14C-label. The tubes were rotated end-over-end at 25 °C. At designated times, the sediment-water slurries were centrifuged for 10 min at 14 000 rpm (Eppendorf 5415 C microcentrifuge). The supernatant was removed, and the remaining sediment plug was washed twice with 10-mL portions of distilled water for 30 min (tubes rotated end-over-end at 25 °C), followed by sequential extraction for 2 h each with 10 mL of 1.0 M NH4OAc and 10 mL of 0.5 M NaOH. The short extraction times ensured that contributions from irreversible sorption during the sorption process were negligible. In each case, the aqueous and sediment phases were separated by centrifugation. One milliliter of each solution, which included the initial aqueous phase, distilled water rinses, and the NH4OAc and NaOH extracts, was then added to 10 mL of scintillation cocktail (Packard, Ultima Gold) and counted for 5 min using a sample channel ratio method (Beckman LS 6000 liquid scintillation counter). Additionally, the concentration of aniline in each solution was measured by LC to determine the amount of 14C-label that could be attributed to the parent compound.

Results and Discussion Initially, the sorption kinetics of aniline and pyridine were measured in a resaturated pond sediment (5% solids) at the nominal concentration of 5 µM. Plots of concentration of aniline and pyridine remaining in the aqueous phase versus time for the Cherokee Park sediment-water system are illustrated in Figure 1. Inspection of the kinetic data indicates that removal of aniline is fast over the first 24-h period of the experiment followed by a slower rate of removal that extended over a 3-week period (Figure 1a). By comparison, removal of pyridine from the aqueous phase was quite fast, reaching equilibrium in 1-2 h (Figure 1b). It should be noted that heat sterilization of the dry sediment was necessary to prevent the rapid loss of aniline that occurred in untreated sediment slurries after a 48-72 h incubation period. This rapid loss was attributed to microbial oxidation of the aniline. A thorough heat sterilization of the dried sediment (3 × 20 min with a 48-h period between heat treatments), however, significantly altered the reactivity of the sediment. Control experiments demonstrated that a single heat treatment had no effect on the sorption kinetics of aniline. Subsequently, heat sterilization of the sediment was limited to one treatment. The biphasic kinetics observed for the loss of aniline in the sediment slurry is a common feature of the reaction kinetics for aromatic amines in soils (2, 6, 9, 13). The biphasic VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2471

FIGURE 1. Aqueous concentration versus time for (A) aniline and (B) pyridine in a Cherokee Park sediment-water system (5% solids, pH 5.5). The initial aqueous concentration of aniline and pyridine was 5 × 10-6 M. Results of duplicate experiments are shown. sorption kinetics for aniline in the Cherokee Park sedimentwater system cannot be adequately described by simple rate expressions. Such an observation is not surprising considering the potential multiple pathways for sorption and the heterogeneity of the reaction system. Sequential Extractions. To provide further insight into the processes controlling the sorption of aniline and pyridine, sequential extractions of sediment treated with 14C-labeled aniline and pyridine were conducted as a function of time to distinguish between reversible and irreversible sorption processes. A plot of % 14C-label versus time for the aqueous phase and each of the extracts is illustrated in Figure 2. Initially, extraction with methanol was performed prior to the salt extraction; however, due to consistently low recoveries of 14C-label ( 4 h.

FIGURE 3. Aqueous concentration of aniline versus time in a Cherokee Park Sediment-water system as a function of pH; pH ) 3.75 (0), pH ) 5.27 (O), pH ) 6.14 (4), pH ) 6.82 (×), pH ) 7.37 (]). solubilized by base extraction. The efficiency for the extraction of soil organic carbon by 0.5 N NaOH has been reported to be in the range of 50-75% (17). Assuming that sorption by cation exchange of the anilinium ion is reversible, the observation that near steady-state conditions are reached after 70 h suggests that there is a limited number of reactive sites for irreversible binding or that access to these sites becomes rate limiting (i.e., a mass transfer limiting process). These results are consistent with those previously reported in the literature for the sorption of aromatic amines to soils. Although the initial sorption of aniline, R-naphthylamine, and benzidine appeared to be dominated by cation-exchange processes in soil slurries, the incomplete recovery of aromatic amine after 24 h suggested that some covalent bond formation occurred on highly reactive sites in the soil matrix (15). In subsequent work, Li and Lee (9) found that reversible sorption processes for aniline and R-naphthylamine in soil slurries dominated at early contact times and that decreasing aqueous solute concentrations over time resulted from irreversible binding/transformation. In a similar fashion, the decrease in the solvent- and NH4OAC-extractability of benzidine, R-naphthylamine, and p-toluidine from soils was accompanied by a gradual increase in NaOH-extractability of 14C-activity providing further evidence for the sorption of aromatic amines to soils through an initial rapid, reversible process followed by a slower, irreversible process involving covalent binding (6). In contrast to the results for aniline, the sorption of pyridine to the sediment can be described solely by a cation exchange process. Unlike aniline, reaction pathways for the covalent binding of pyridine do not exist. The extremely fast sorption kinetics for pyridine are consistent with cation exchange, which is typically described as an instantaneous process whose rates are diffusion controlled (18, 19). pH Effects. Speciation of aniline over environmentally significant pHs requires an understanding of how pH affects the competition between reversible and irreversible sorption. Toward this goal, the sorption kinetics for aniline were measured in a Cherokee Park sediment-water system as a function of pH. Solution pH was maintained by the addition of appropriate buffers. The concentration versus time plots are illustrated in Figure 3. Table 2 provides a summary of the speciation of aniline as a function of the pH of the sediment slurry, the fraction of aniline sorbed at t ) 4 h and the pseudofirst-order rate constants calculated from the concentration data at t > 4 h. These data demonstrate that (1) the initial sorption of aniline (t < 4 h) is pH dependent (increasing sorption with decreasing pH), and (2) the longer term sorption rate (t > 4 h) is nearly independent of solution pH above the pKa of aniline (4.64) and increases by only a factor of 2 at solution pH below the pKa of aniline.

The effect of pH on the initial sorption of aniline is readily explained as a shift in the solution equilibrium toward the protonated species with decreasing pH and a subsequent rapid removal of the protonated species through a cation exchange process. The pKa of aniline is 4.64, thus over the solution pH range studied (3.75 to 7.37), the speciation of aniline is dramatically altered (Table 2). At pHs 6.82 and 7.37, where less than 1% of the aniline is expected to be in its protonated form, the fraction of aniline sorbed at t ) 4 h is 0.18 and 0.16, respectively. As the solution pH is dropped to 6.14 and 5.27, and subsequently to 3.75, which is below the pKa of aniline (4.64), there is an increase in fraction of aniline in its protonated form as well as the fraction of aniline sorbed at 4 h. At pH 3.75, where the fraction of protonated aniline is 0.89, 70% of the aniline is sorbed within 4 h. At pH values above the pKa of aniline (4.64), the pseudo-first-order rate constants for the slower rate of disappearance of aniline varied by less than a factor of 2 (Table 2). At pH 3.75, below the pKa of aniline, the rate constant increases by a factor of ∼2. Although it is the neutral form of aniline that is the reactive species involved in reactions leading to irreversible sorption, the decrease in the aqueous phase concentration of aniline due to enhanced cation exchange does not impede covalent binding as a result of decreasing the aniline concentration in solution. Effect of Aniline Concentration on Sorption Kinetics. To provide further insight into the processes controlling the sorption of aniline in the sediment slurries, sorption kinetics were measured as a function of initial aniline concentration. The initial aqueous concentration of aniline was varied by more than 2 orders of magnitude, ranging from 5 µM to 1 mM. The experimental results are illustrated in Figure 4 as a plot of aniline concentration (normalized to the initial concentration) versus time as well as a plot of sorbed aniline concentration (Cs) versus aqueous phase aniline concentration (Caq) at selected times. Sorbed aniline concentrations were determined by difference from the total and the aqueous phase aniline concentrations. These data clearly demonstrate that the sorption capacity of sediment significantly exceeds that first indicated from experiments with aniline at the nominal aqueous concentration of 5 µM. Even at the highest initial aqueous concentration investigated (1 mM), it is apparent that slow removal of aniline from the aqueous phase continues after 500 h. The Freundlich parameters (Table 3) for the sorption isotherms in Figure 5b show that the Freundlich constant and the nonlinearity of the isotherms increase with time. In a separate experiment, Cherokee Park sediment slurries treated with aniline at an initial concentration of 0.005, 0.1, and 1.0 mM, respectively, were extracted with 1.0 M NH4OAc after a 240 h incubation period to determine if the distribution between reversible and irreversible sorption varied with aniline concentration. At the nominal initial concentration of 0.005 M, less than 5% of the sorbed aniline could be recovered by treatment with 1.0 M NH4OAc VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2473

FIGURE 4. (A) Aqueous concentration of aniline normalized to the initial concentration versus time (5% solids, pH 5.5); 5.10 × 10-6 M ([), 1.02 × 10-5 M (0), 1.02 × 10-4 M (2), 5.11 × 10-4 M (]), and 1.02 × 10-3 M (O). (B) Freundlich sorption isotherms as a function of time.

TABLE 3. Freundlich Parameters for the Aniline Sorption Isotherms as a Function of Time t (h)

Kfa

Nb

r2 c

24 96 216 528

34.6 41.4 48.7 57.7

0.65 0.58 0.54 0.50

0.994 0.994 0.995 0.992

a Freundlich constant isotherm (Cs ) KfCNaq).

b

Measure of nonlinearity

c

Fit to the Freudlich

(consistent with data set in Figure 2). At the initial aniline concentration of 0.1 mM, 26% of the sorbed aniline could be recovered by extraction with 1.0 M NH4OAc after 240 h, whereas at 1.0 M initial aniline concentration, 55% of the sorbed aniline could be recovered by extraction with NH4OAc. The observation that the fraction of reversibly bound aniline (i.e., fraction that could be recovered by extraction with 1.0 M NH4OAc) increases with increasing initial aniline concentration suggests that there is a limitation to the reactivity and/or availability of irreversible binding sites (see subsequent discussion). At the nominal initial concentration of 0.005 M, the pool of aniline available for reversible sorption is depleted by chemical processes resulting in irreversible sorption of aniline. This occurs to a lesser extent at the higher initial concentrations of aniline. In subsequent experiments, the availability of sorption sites as a function of time was determined by respiking Cherokee Park sediment slurries with aniline after an initial 175-h incubation period. These studies were conducted at an initial concentration of aniline of 5 µM and 50 µM, respectively (Figure 5). Although the sorption kinetics after respiking with aniline were somewhat slower than those observed for the initial addition of aniline at the nominal aniline concentration of 5 µM, these data demonstrate that a portion of the most reactive sorption sites in the sediment 2474

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 12, 2001

FIGURE 5. Effect of respiking with aniline on sorption kinetics in a Cherokee Park sediment-water system (5% solids, pH 5.5). At t ) 175 min, the aqueous aniline concentration was increased to the initial aqueous concentration of (A) 5 × 10-6 M and (B) 5 × 10-5 M. matrix were still available. Three hundred hours after respiking the sediment slurry, the aqueous phase concentration of aniline approached that just prior to respiking. By contrast, at the higher initial aniline concentration of 50 µM, we no longer observed biphasic kinetics for the disappearance of aniline upon respiking the sediment slurry after the 175 incubation period. Apparently, most of the highly reactive sorptive sites (i.e., cation exchange sites and fast reacting covalent binding sites) and/or the readily accessible sorptive sites (surface associated) were consumed by the initial treatment of aniline. In a similar fashion, about 10% of the covalent binding sites associated with a river fulvic acid isolate were highly reactive and quickly became saturated by treatment with aniline (8). Mass-Transfer Limited Covalent Binding. A common feature of the sorption kinetics observed in this study for aniline as well as other sorption studies of aromatic amines in soil and sediment slurries is the slow approach to sorption equilibrium. This same type of behavior is commonly observed for the sorption of hydrophobic organic chemicals (HOCs) and is attributed to mass-transfer limitations (ratelimiting molecular diffusion) (20, 21). Generally, the slow sorption kinetics for aromatic amines have been attributed to a rate-limiting reaction between the aromatic amine and specific sites (e.g., electrophilic carbonyl groups) on the soil surface (2, 6, 9, 13). We suggest, however, that mass transfer limited processes will also play a role in the nonequilibrium sorption of aromatic amines. The effect of initial aniline concentration on sorption kinetics are consistent with the scenario in which the slow, sorption kinetics of aniline are limited by the reactivity and/or availability of covalent binding sites. Based on these data we cannot distinguish between mechanisms for slow sorption that are limited by slow chemical kinetics or mass transfer limited processes. More recent studies with substituted anilines in Cherokee Park sediment, however, have provided further insight into the

processes controlling the slow, sorption kinetics of aniline (22). The rate constants for the initial fast sorption for a series of monosubstituted anilines correlate well with molecular parameters describing the nucleophilic reactivity of the amino group (e.g., pKa’s and σ-constants). By contrast, the rate constants for the slow sorption kinetics of substituted anilines do not correlate well with these same molecular parameters suggesting that rate-determining processes other than chemical kinetics are controlling the slow sorption kinetics of the aromatic amines. We can hypothesize as to the mechanisms for the mass transfer limitations that may contribute to the slow, irreversible sorption of aniline. Mechanisms invoked for the slow sorption kinetics of HOCs include rate-limiting diffusion through bulk and film water, pore diffusion, and matrix diffusion (20, 21, 23). We consider diffusion through bulk water and film water to be insignificant because the sediment slurries were well mixed and the sediment consisted of small particle size (